There has been a huge increase in e-learning courses, and the Recommendation systems (RSs) are useful tools for narrowing down the course options and expose students to the courses that suit their needs. The majority of the research related to recommendation systems focus on effectiveness rather than factors influencing its acceptability, and in practice, user satisfaction cannot be explained by accuracy alone. This study will consider course recommendation system and investigate the acceptance of the courses suggested by the recommender system (RS) to scholar’s or research students based on their learning needs which in turn will help to find out why some individuals accept new technology while others resist. Therefore, research scholars (n = 150) who willingly engaged in this study were asked to use the RS and complete a questionnaire also based on their experience as part of a self-administrated longitudinal survey. This study assessed external factors such as perceived availability, relevance, and experience, which were not explained by the Technology Acceptance Model (TAM). The results confirm that the extended TAM provides a valuable theoretical model, which helps to understand scholar’s acceptance of RS, and other factors positively that affect the original variables of TAM. Hence, a new modified TAM that incorporates three external factors has been offered. The considerable value calculated for Cronbach’s alpha confirms that the results are valid and reliable. The observed results will help recommendation system developers to maximize user experience because the implications of this research effort are critical for instructors, scholar’s, and institutions as well.
Research Article
Integrating External Factors and Technology Acceptance Model to Understand Scholar Intention and Use of Recommendation System for Course Selection.
https://doi.org/10.21203/rs.3.rs-2126671/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Recommendation System
Technology Acceptance Model
information Retrieval
Cronbach's alpha
Course Selection
The popularity and advancement of web technologies has resulted in a plethora of advanced applications that can recommend alternative things to users based on their requirements.Various possibilities arise which stress the need for the usage of the Recommender system (RS) and the way it can be integrated into education. As a result, it became indispensable to assess the use of the Recommender system in a different way. Therefore, RS have been seen as a valuable tool where it has been used to make recommendations for news, movies, music, and online connections, among other things.
A recommendation system is a tool that uses different information retrieval techniques to filter the Items or learning resources and then present them to the user (Chughtai, Selamat, Ghani, & Jung, 2014). The recommendation systems have been developed across different domains such as News, Music, Movies, and e-commerce (Dascalu et al., 2016), e-tourism, e-governance, (Lu, Wu, Mao, Wang, & Zhang, 2015), where it has achieved a huge success rate. The Recommender systems are very popular areas of research that provides need based suggestions (Adomavicius&Tuzhilin, 2005), but the algorithms underlying traditional Recommendation systemssuch as News, Music, Movies, and e-commerce, cannot be redeemed in educational systems, as recommendations in the educational scenario may change over a while due to change in the students need or the context (Núñez-Valdéz et al., 2012).
The main chore of the Recommender system in e-learning systems is to recommend the appropriate courses or the learning material that is relevant concerning Student needs and helps them in making decisions (Aguilar, Valdiviezo-Díaz, &Riofrio, 2017). It is logical that applications of Recommender system should be made available for the learners and to call for the effectual acceptance of recommendations based on relationship, preferences, and attributes among Students (Ricci et al., 2011).Many research focused on the use and acceptance of information technology. In the literature, recommendation systems have a significant impact on emerging information system techniques, and they usually centred on individual experience and skills for assessing the potentially large quantity of items that are available (Selim, 2003). Many researchers are currently investigating the elements having a direct impression on acceptance of recommendation systemby its users for maximizing its recognition. Malaysian research group (Ahmed et al., 2011; Malathi and Rohani, 2011; Hussein, 2017) investigated the impact of new technology by studied factors like its usefulness,intention to use, ease of use, understanding user’s preferences, the intent of reusethe system,accuracy in recommending items, and interaction with the system.
Anderson, D. L. (2004) using a learner-centred approach tried to obseve the significance of meeting online Student needs. Hughes (2004) on the other recommends creating communities of online learners to provide online support services to meet the needs of online students and for enhancing the interaction among Students. Zhu Z, Yu M, Riezebos P, et al., (2016) find out that in this age of Information technology the concept of smart education is to portray a kind of new learning process. Burdea&Coiffet (2003) believes that the latest developed technologies will help in seamless personalization and assist learners to learn in a more flexible, effective, and comfortable way. There are high chances that the outcome of this progressive shift in technology will bring more prevailing and interactive modes of communication and ensure the delivery of learning content of high quality.
The efficiency of the recommendation systems, mainly, dependent on the factors that are ahead of the prominence of the recommendation algorithms. To address the difficulties of forecasting and describing system utilisation, many descriptive models in information systems have been investigated.The Technology Acceptance Model (TAM), on the other hand, stands out and has widespread acceptance in the information systems field. (Swearingen and Sinha (2001), Chuttur, 2009).The idea of reasoned action provided by (Fishbein and Ajzen, 1980) was used to develop this model TAM model is based on user beliefs, attitudes, and intentions in technological adoptions (Davis et al., 1993).There are two more criteria that constitute the construct of belief: perceived usefulness andperceived ease of use. Both of these determinants, as well as intention and attitude, constitute a sequence that illustrates the user's system adoption.Hsiu-Mei Huang and Shu-Sheng Liaw (2018) showed that perceived interaction and perceived self-efficacy are two essential factors, which affect learning motivation, perceived usefulness, and perceived ease of use. Authors also claimed that motivation towards learning is also an interpreter to affect perceived usefulness.
This study is conducted keeping course recommendations as a test-bed and a questionnaire appropriate for an educational domain. In the form of the TAM model, three new external variables will be introduced to the baseline: lack of RS availability, experience, and relevance. The goal of this study is to see how well university students especially research scholar’s embrace recommender systems, and to learn more about the factors and underlying acceptance and behavioural intention to use the recommendation system for course selection. To our knowledge, there are no technology acceptance studies connected to the recommendation approaches used in educational domain or in e-learning. To fill this research gap, this study looks into the perceived availability, relevance, and experience of the recommender approach, as well as the actual deployment of recommendation approach in such models. The remainder of the document is organized as follows. The aspects of the Technology Acceptance Model employed in this study are briefly explained in the next section. In Section 3 the associated work concerning TAM applications of recommendation system will be discussed. In Section 4, research model is proposed and in Section 5 the chosen methodology is described. Section 6 discusses the results analysed with the proposed model, and Section 7, 8 presents the discussion and conclusion respectively.
Most of the higher educational institutions have adopted e-learning systems and educational technologies (Potkonjak et al., 2016).Higher education institutions can no longer afford to fall behind in terms of technological breakthroughs in teaching and learning. Thus, the knowledge that learning technologies might revamp higher educational facilities is attributed to their quick development. (Farrell &Rushby, 2016). Gil A B, Garciapenalvo F J. (2008) believed that the selection of courses will play an important role in the development of technology-based education which allows Students to take courses that are suitable for them with more accuracy and professionalism. Jinjiao Lina et al., (2018) Introduced a Framework based on the Sparse linear method (SLIM) for generating top-n course recommendations which are appropriate to Students by adding both sparseness regularization and expert knowledge in the computation. Load of information in an E-learning environment has increased as thus the burden of information has increased manifolds, resulting in various difficulties for Students as they struggle to choose a course of interest.Nowadays, recommender systems handle information loads in E-commerce (Hubert et al., 2017), online libraries, E-health, Sports, and, to a lesser extent, education by customising and filtering data according to the needs of Students (Zameer et al., 2018). Personalization offers the added benefit of guiding Students through the course selection process based on their interests and needs.Personalization in the form of suggestions has gained traction and is being implemented in a variety of domains, including social, business, and health (Hazra et al., 2014). Recommender systems are helpful in assisting users in making informed selections with minimal effort. However, any new technological notion, no matter how simple or complex, is subject to acceptance testing. Various models have been developed to characterise the relationship between technology and its users. A very general concept is given in Fig.1.
There is a lot of study on user acceptability of technology, and TAM is one of the most well-known. It is being accepted because of its ease of use, robustness, and applicability in explaining and predicting the elements that influence user adoption of new technologies. (Rauniar et al., 2014;Marangunić&Granić, 2015).To evaluate recommendation systems acceptance by users' various researchers have tested TAM for different purposes. Mobile shopping (Hubert, Blut, Brock, Backhaus, &Eberhardt, 2017),retail and banking industry for evaluating Recommender system adoption (Asosheh et al., (2008)). Similarly (Hu & Pu, 2009)evaluate the Recommendation system that is based on personality using TAM. Their main goal was to propose music based on the user's personality as well as their emotions and mood. Based on the behavioural parameters mentioned in TAM, another researcher (Lee et al., 2007) built a system for recommending virtual communities to an active user. Filtering functions based on user needs were added to the recommendation process, and a customised version of TAM was used to assess individual approval of the recommender system for an online trading experience. (Baier& E. Stüber, 2010).
One of the most important foci of studies is the adoption of technology by intention and the use of independent factors. As a result, a number of variables are involved in the acceptability of recommendation technology in terms of its users' experience with the system, as it is a complex process that precedes having a visually appealing and simple-to-use interface. Even similar interface systems have been discovered to be perceived differently by consumers when the algorithm underpinning them is modified. (Armentano et al, 2014). (Davis, 1989) attempted to answer issues about what factors influence users' acceptance or rejection of newly discovered information technology. Initially, it was determined that users would utilise an application if they believed it would be useful and perform better than if they did the task without it. It was also discovered that, even if an application is difficult to use, if users believe it is valuable, the perceived benefits of using it outweigh the effort required. Both of these findings were labelled as factors of "perceived usefulness" and "perceived ease of use" in the famous Technology Acceptance Model (TAM).
TAM and Extend-TAM, also known as TAM2, have proven to be quite useful in research projects. Many research papers, on the other hand, have debated and negotiated the flaws and weaknesses revealed in these models. The following research studies provide a list of those problems. (Abidi& Khan, 2020; Dwivedi& Rana, 2019). The TAM-2 by (Venkatesh & Davis, 2000) whose focuses on technology in place of work, proposed an external variable output that affects perceived usefulness. But the output for a job will be high only if there is some job relevance that will be seen by the users and without relevance the chances of accepting new technology are low. In our proposed model, relevance has been taken as one of the external factors to learn its effect on technology acceptance. Even Senior Technology Acceptance or STAM (Renaud & Van Biljon, 2008) tried to check individual factors of differences for accepting technology usage. But, they have not included the availability feature to learn about the individual differences for accepting the latest technology. If availability will not be accounted then it’s very difficult to talk about the individual differences. The third variable experience has been very much highlighted by (Leong, Ibrahim, Dalvi-Esfahani, Shahbazi, &Nilashi, 2018; Liébana et al, 2016) to facilitative the role it played in technology adoption and they found that the influence of Experience on Perceived Usefulness is incompatible. But they have not studied the role of experience concerning user behavioural intentions and attitude towards using the new technology which will be the focus of this study. Hye Lee and Stoel (2003) found out that the user’s experience with the technology goes up, in such cases the chances of perceiving the technology are more. Therefore, there is an evident relationship between of relevance of technology, its availability, and the experience of its usage among the users, which needed to be studied.
This research work sought to understand the acceptance of Recommender systems by Indian research scholars by modelling the relationship between the original TAM and external constructs. The questionnaire was developed for testing the hypothesis as O’Leary (2017) recognized several advantages of this research methodology and suggested adapting the existing available questionnaire. Thus, we have developed the questionnaire by following the guidelines and criteria followed by other models. Based on earlier research and theoretical concept, this study extended TAM in recommender system application by including three external variables i.e. Relevance, Availability, Experience. The dependent variable for this study is behavioural intention, whereas the rest of the constructs are independent variables. In Fig.2 the graphical representation of proposed research model with hypothesized relationships created for relevant constructs.
A new theoretical model was developed with the help of the earlier research theoretical model. The core constructs of the TAM will be used in the research model described here: perceived ease of use, perceived usefulness, behavioural intention to use, and actual system use. The model also includes three more variables: Availability, Relevance, and Experience are all lacking. Based on their expertise choosing courses and the availability of technology for such tasks, the research model is aimed at research researchers.As previously suggested (Taylor and Todd, 1995), the behavioural intention to use technology is almost identical across inexperienced and experienced users, and TAM can be applied even before the adoption of new technology (Shih, 2004). This research is significant because it considers a variety of factors. To begin with, no previous research has looked into the availability, experience, and relevance of Recommender systems in relation to behavioural intention to use them and validated it with TAM. This discovery also sets the path for future research into improving E-learning systems. As a result, the questionnaire utilised in this study could be used in future studies.
3.1. Research hypothesis
a) Hypotheses in relation to TAM variables
Various studies and RS usage have verified the relationship between the various TAM constructs.(Zouhair, 2012). As per Venkatesh and Davis (2000), PEU (Perceived Ease of Use) is linked with an individual’s observation concerning how the new technology will be easy to use. Therefore, students (Featherman&Pavlou, 2003) measure it as an activator for adopting advanced technology. Perceived Usefulness is supposed to be the main variable to drive usage behaviour (Chen, Cheung, &Lee2005).Legris et al. (2003) in their research found that 80% of the studies based on TAM have shown the impact of Perceived Ease of Use on Perceived Usefulness & Perceived Usefulness on Behavioural Intention.
Therefore, the connection between perceived ease of use, perceived usefulness, attitude toward using, and intention to use RS system is hypothesized as the following
H-1: There is a positive effect of Perceived usefulness on attitudes towards using RS.
H-2: There is a positive effect of Perceived usefulness on behavioural intention of using RS.
H-3: There is a positive effect Perceived ease of use on perceived usefulness of RS.
H-4: There is a positive effect Perceived ease of use on attitude towards using RS.
H-5: There is a positive effect Perceived ease of use on intention to use RS.
H-6: There is a positive effect Attitude towards using on the user’s intention to use RS.
b) Hypotheses with respect to TAM variables and External factors
After evaluating the studies, the TAM constructs usefulness, it is possible that ease of use is insufficient, and additional external variables may be included in the study(Ronnie, 2011). Experience has a positive effect on Perceived Usefulness (Yanxia&Xiangling, 2019). Some of the external variables used in this study are RS availability, RS relevance, and RS Experience. External variables are shown in Figure.2, as we believe that these variableswill affect the original TAM since such factors have not been studied at all in research as discussed in the literature survey. The hypothesis is as follows:-
H-7: RS Experience can have a positive and significant effect on attitude towards using.
H-8: RS Experience can have a negative effect on behavioural intention of use.
H-9: RS Relevance can have a negative effect on perceived ease of use.
H-10: RS Relevance can have a positive and significant effect on perceived usefulness.
H-11: RS Availability can have a positive and significant effect on attitudes towards use.
H-12: RS Availability can have a positive and significant effect on perceived usefulness.
4.1 Participants
This research work used the convenience-sampling technique for data collection in which 150 Ph.D. scholars of computer science were given a trial of the system where they have provided the research title as queries as shown in Table 1to our system and in return the system has recommended some courses to them as shown in Figure 3 and Figure 4 , along with the ontologyand inference based diagram in Figure 5and Figure 6 to make them understand the hierarchy of the subject in a particular to area. Then they were asked to give their feedback with the help of a questionnaire. The online survey would have been an appropriate tool because the demonstration was an important task to be carried out to receive genuine feedback from scholars. The scholars who participated in the experiment were from computer science engineering, Information technology, computer applications, and mathematics department. There were 69 Male, 81 Females in the age group of 24 to 46 years and they are generally aware of the traditional way of choosing the course work for their Ph.D. program which could guarantee their evaluations of using the Recommender system in a much better way.
To ensure the content's validity, a questionnaire was created in this study based on the fundamental TAM assessment scales and other literature (Sánchez and Hueros, 2010), with slight modifications to meet the context of the Recommender system. The survey was established to investigate the relationship between constructs in the suggested research model, and the experiment was largely conducted with research scholars (Ph.D.) and faculty (perusing PhD) from institutes of Tamilnadu, where we introduced our course recommender system to them. We evaluated our research model for another reason: learners have looked into traditional course selection methods, thus this study allowed us to compare previous outcomes.
Table.1 Research Queries given by Scholars to our Recommendation System
|
Scholars Research Title as Query |
Relevant Courses by our Keyword based system |
Most Relevant Courses by Our Hybrid Recommendation System |
|
Knowledge discovery from databases using AI |
2 |
4 |
|
Data Analytics in Healthcare System |
2 |
4
|
|
Software Requirements Elicitation Techniques |
2 |
5 |
|
Knowledge discovery in E-learning Systems |
2 |
5
|
|
Malware Detection And Protection Against Web Attacks |
3 |
5 |
|
Opinion Mining And Social Interaction Using Clustering Algorithms. |
2 |
6 |
|
Opinion Mining in Online Customers Reviews Using Supervised Method |
2 |
5 |
|
Energy efficient resource management in cloud computing |
2
|
5 |
|
Network Security Analysis and Detection of Ports Scans |
3 |
5 |
|
An Effective Hybrid Gateway Discovery Scheme in Integration of Internet & MANET |
2
|
5 |
|
Biometrics Security & Forensics in digital passport system |
3 |
5 |
|
CBIR Using Image Mining Techniques |
1 |
4
|
|
Location Based Web Targeted Advertisement on Cloud |
2 |
5 |
|
Knowledge Extraction from Social Web Source for Online Recommendation System |
3 |
5 |
|
Study on the effectiveness of Agile methods on Communication product software engineering |
3 |
6 |
|
Average Relevant Courses |
2.2 % |
4.93 %
|
4.2 Instrumentation
It's difficult to get a response from the complete population; sampling is the only way to draw conclusions from a limited sample size in a particular community (Jemain et al, 2007).The survey sample consists of researchers and research faculty members from three different universities, and convenience sampling (Non-Probability) was utilised because it has been used in many studies on technological acceptability due to its shorter time period and higher response rate. As a result, convenience sampling proved to be the most effective strategy for this study.Section 1 of the research instrumentation deals with the demographic information, experience, availability, and relevance of the participants.Section 2-5, on the other hand, employs a five-point -Likert response scale. For our new external variable, ten (10) questions were multiple choices, while the rest were developed using the Likert-5 scale with matching options: strongly Agree (1), Agree (2), can't say (3), Disagree (4), Strongly Disagree (5). The questions posed, as well as the TAM variable linked with them, are detailed below. To ensure lucidity, unambiguousness, and comprehensiveness, three teachers with experience of handling research scholars were consulted for doing the pilot test & seven scholars who were in the stage of completing their PhDs to check the readability of the content present in the survey tool. The final version consisting of 30 items (excluding Name and Gender) that were presented to the 150 participants.
4.3. Data collection
Over 175 people had access to the questionnaire that had been produced. The response percentage reported was 85.71 percent, with 150 valid responses afterwards used for analysis. External variables are measured in the first section of the study, while TAM components are measured in the second section, as mentioned in the questionnaire design above. As per the thumb rule, we should have a 5:1 ratio of participation concerning the total number of valid items (Bentler& Chou, 1987). While the minimum sample size for the estimation of likelihood should be 100 as per the recommendations given by (Hair, Anderson, Tatham, & Black, 1998). In this study, the questionnaire comprises 30 items which give a suitable sample size of 150. Similar sample size results were obtained by previous studies (Rupp et al., 2018; Haapasalo&Kadijevich2008), therefore collective size of sample (n=150) of this study is adequately effective and the details are provided in Table 2.
Table 2.Participants Characteristic
|
Attribute |
Variable |
Frequency |
Rate |
|
Gender |
Male Female |
37--- 69 43 ---81 |
46.% 54% |
|
Age |
23-30 31-40 > 40 |
96 51 3
|
63% 34% 3% |
|
Time Spend on Course Selection |
1-2 weeks 3-4 5-8 > 8 |
12 84 45 9 |
8% 56% 30% 6% |
|
Course Satisfaction & Relevancy |
Fully Partial Irrelevant |
12 108 30 |
8% 72% 20% |
|
Designation |
Scholars Faculty |
115 35
|
76.% 24.% |
|
Course Selection Methodology |
Manual Tech. Assisted |
150 Nil |
100% 0.0% |
5.1 Statistical Analysis
Table 3. Statistical Analysis.
|
Constructs |
Mean |
Standard Deviation |
Kurtosis |
Skewness |
|
Perceived Ease of Use (PEOU) |
4.53 |
0.61 |
2.9 |
-0.93 |
|
Perceived Usefulness (PU) |
4.45 |
0.58 |
2.3 |
-0.54 |
|
Attitude towards use (ATU) |
4.44 |
0.62 |
2.54 |
-0.67 |
|
Availability |
2.14 |
0.75 |
1.91 |
-0.25 |
|
Relevance |
2.14 |
0.44 |
4.3 |
0.64 |
|
Experience |
1.80 |
0.35 |
3.58 |
-1.52 |
In Table 3, Mean, Standard deviation, Skewness, and Kurtosis (as data normality) has been provided for PEOU, PU, ATU, Availability, Relevance, and Experience. The internal constructs were given a value range from 1-5, with 3 being the midpoint and the external constructs were given range from 1-4 with 2 being the middle point, as the number of items in external constructs is less compared to internal constructs. The range of mean is 4.44-4.53 for the first three constructs, which indicates that the overall response of scholars towards the recommender system is positive. The mean value further implies that the scholars have perceived more ease of use (PEOU=4.53) than the perceived usefulness (PU=4.45) and Attitude towards using (ATU=4.44) the recommender system. The range for standard deviation is between 0.354 to 0.75, suggesting that the feedback of the scholars are barely stretched. For checking the normality of the data skewness, the bounded likelihood estimation value as per Kline (2005) and Wong (2016), should be within +3 and for Kurtosis, the range should be under+ 10. Table 3 shows the desired value range for data normality for all the items in both Skewness and Kurtosis.
5.2 Validity and Reliability Analysis
The two-step approach suggested by (Anderson and Gerbing (1988)) analyzed the model by assessing validating the reliability of the variables and SEM was used to check the variable significance. To evaluate the constructs validity and reliability, the Average Variance Extracted (AVE) and composite reliability were computed and initially discriminant validity, reliability and convergent validityof the constructs were also tested.Within a single factor, reliability is used to assess the consistency of item-level mistakes. Additional tests were conducted utilising Cronbach's alpha reliability assessment in addition to the previously described reliability and validity tests to examine instrument stages. Cronbach's alpha is internal consistency metric or a scale reliability that indicates how consistently a set of dependable variables loads on the same factor. (Hair et al, 2006). Cronbach's alpha is a function of the number of test items and the inter-correlation (average) between them. This reliability metric will assess how closely a bunch of objects are linked. When the value of Cronbach's alpha reaches 0.07, constructs are regarded to have internal consistency dependability, according to studies. (Sekaran, 2006). By using SPSS 21.0, coefficients of Cronbach's alpha wereacquired to evaluate the relationships in the structural model. All of the metrics in this study have a good level of reliability, ranging from 0.755 to 0.863, with 0.863 being a satisfactory result for both PU and Relevance. The average scale was greater than 0.70, indicating that the survey was reliable. Table 4 shows the Cronbach's alpha coefficient for several criteria. Cronbach's-alpha is greater than 0.7 for all five components, indicating that the data is reliable. This value can be deemed adequate or sufficient in this study. By using the FL > 0.5, (factor loading in Eq.1) convergent validity was tested, for CR > 0.7 (Composite Reliability in Eq.2) and AVE > 0.5 (Average Variance Extracted in Eq.3) (Cheung, R., & Vogel, D. (2013) similar tests were done. The results in Table 4 suggested that the value of FL is above 0.5 ranged from 0.778- 0.894. It also shows all-composite reliability measures ranged from 0.76 to 0.923.
|
ei = 1- λi2 Eq.(1)
|
The standardised factor loading for item i is (Lambda), while ε is the respective error variance for item i. As illustrated in equation Eq.2, the error variance ( ε) is calculated using the value of the standardised loading ((λ). while as Average variance is given in Eq.3.
Table 4.Convergent Validity and Reliability of the Constructs.
|
Constructs |
Items |
Cronbach's alpha |
Factor loadings |
CR |
AVE |
|
Perceived Ease of Use (PEOU)
|
PEOU1 PEOU2 PEOU3 |
0.765 |
0.894 0.845 0.848 |
0.920 |
0.75 |
|
Perceived Usefulness (PU) |
PU1 PU2 |
0.863 |
0.838 0.877 |
0.765 |
0.53 |
|
Attitude towards use (ATU) |
ATU1
|
0.755 |
0.857 |
0.923 |
0.75 |
|
Behavioural Intention to USe (BIU) |
BIU1 |
0.890 |
0.894 |
0.923 |
0.750 |
|
Availability
|
A1, A2 |
0.764 |
0.787 0.787 |
0.76 |
0.61 |
|
Experience
|
E1,E2 |
0.763 |
0.794 0.778 |
0.76 |
0.76 |
|
Relevance
|
R1, R2 |
0.863 |
0.887 0.857 |
0.86 |
0.61 |
As a result, the suggested model has met the item reliability recommendation. Convergent validity is established when different items are utilised to assess the same construct and the findings from the different items are substantially related. Table 4 further demonstrates that each item's factor loadings are highly significant (p<0.001) and greater than 0.5. If the constructs can be separated sufficiently from one another, discriminant validity is assessed, and if the square root of the AVE for a construct is greater than its correlations with other constructs, same validity is established. (Fornell&Larcker DF, 1981). The value of CR > 0.7is above the satisfactory measures and the values of AVE greater than 0.5 is also above the acceptable criteria. The overall results indicated that convergent validity for all the constructs was satisfactory. Table 5 shows that the model meets the requirement for discriminant validity efficiently. We have seen that all appropriate reliability measures and fit indexes come within the suggested ranges; signifying the fact that the measurement model fulfilled all criteria for model fit, construct validity, and reliability. As a result, the model might be used to evaluate the hypotheses in Section 3 about causal relationships.
Table 5. Correlations as Discriminant Validity (Square Root of AVE in Diagonals).
|
Constructs |
PEOU |
PU |
ATU |
Availability |
Experience |
Relevance |
|
PEOU |
0.866 |
- |
- |
- |
- |
- |
|
PU |
0.687** |
0.729 |
- |
- |
- |
- |
|
ATU |
.812** |
.835** |
0.867 |
- |
- |
- |
|
Availability |
0.298*
|
0.332*
|
0.427* |
0.786 |
- |
- |
|
Experience |
-0.198 |
0.278* |
0.324* |
0.331* |
0.786 |
- |
|
Relevance |
0.212* |
0.298.* |
0.321* |
0.327* |
0.363* |
0.872 |
Note: ***P <0.001, **p < 0.01, *p <0.05, Square roots of AVEs values are presented diagonally
In order to ascertain a bi-variate relationship between the variables or constructs the analysis for finding the correlation was conducted. Cohen (1992) suggested that the relation effect could be estimated in terms of strong, medium and small if and only if the range will fall between 0.5 < r < 1.0, 0.3 < r < 0.5 , 0.1 < r < 0.3 respectively. In Table 5,except for the correlation between Experience and Perceived Ease of Use, allother correlations are statistically significant a positive. The strong positive association is found between Perceived Usefulness and Perceived Ease of Use, Attitude Towards Using and Perceived Ease of Use PEOU - Perceived Usefulness. The medium type of association is found between Availability and Perceived Usefulness- Attitude towards Using, Experience and Availability - Attitude towardsUsing, Relevance and Experience –Availability, Attitude Towards Using and Perceived Usefulness. The higher and medium values of correlation coefficients among external variables suggest that for path analysis these external variables can be considered (Cacciamani et al., 2018), and that is the reason they were carefully in the part of path analysis.
For discriminant validation the values of square root of Average Variance Extracted (AVE) was tested to check co-relationship between constructs and the relations lower than Sqrt of AVEwill be easily confirmed for discriminant validity (Fornell&Larcker, 1981, Esteban-Millat et al., 2018). The diagonal values of co-relation i.e the square roots of AVE in Table 5, shows that they are greater than correlations values between the constructs and thus can be considered discriminant validity for all constructs. This in-turn suggests that all the variables discussed in the model exhibit higher discriminant validity, validity of convergence along with adequate reliability.
5.3 Structural Model
Theproposed modified models fit indices final summarization is shownin Table 6. When the entire indices falls in the literature-recommended value ranges, the constructs are said to be well-fit. The data include the Goodness-of-Fit Index (GFI), Adjusted Goodness-of-Fit Index (AGFI), Comparative Fit Index (CFI), Root Mean Squared Residual (RMSR), and Root Mean Square Error of Approximation (RMSEA). After determining the measurement model's reliability and validity, we performed analysis of the path to investigate the association between the latent variables.AMOS 23.0 was used to test the structural model.The constructs are well-fit because entire indices fall in the range of literature-recommended values. Figure 3 depicts the outcomes of the created path analysis.Except for H10 and H7, the paths from Experience to Attitude toward Using and Relevance to Perceived Usefulness. All of the path coefficients in Table 7 were determined to be statistically significant. After taking into account the correlations, estimated pathways, and modification indices, the inconsequential pathswere eliminated,and the paths which are significant were taken into account.
Table 6. Summary of structural model fit indices
|
Fit Index |
Critical Value |
Measurement Model |
Explanation |
|
X2 / df |
<3 (Kline, 2005) |
1.35 |
Good |
|
Goodness of Fit Index (GFI) |
> 0.90(Kline, 2005) |
0.934 |
Good |
|
Comparative Fit Index (CFI) |
> 0.95(Kline, 2005) |
0.967 |
Good |
|
Adjusted Goodness-of-Fit Index (AGFI) |
> 0.90(Kline, 2005) |
0.916 |
Good |
|
Root Mean Squared Residual (RMSR), |
< 0.10(Byrne, 2013) |
0.058 |
Good |
|
Root Mean Square Error of Approximation (RMSEA) |
< 0.08(Byrne, 2013) |
0.03 |
Good |
|
Normed Fit Index (NFI) |
> 0.90(Byrne, 2013) |
0.901 |
Good |
Table 7. Parameter Estimate of path analysis
|
Path |
Standard Coefficients |
Standard Error |
Supported |
|
H1 : ATU |
0.41*** |
0.18 |
YES |
|
H2 : BIUPU |
0.47*** |
0.17 |
Yes |
|
H3 : PU |
0.35*** |
0.15 |
Yes |
|
H4 :ATU |
0.42*** |
0.17 |
Yes |
|
H5 : BIU |
0.48*** |
0.14 |
Yes |
|
H6 : BIU |
0.45*** |
0.16 |
Yes |
|
H7 : ATU |
0.08 |
0.10 |
No |
|
H8 : BIU |
0.57*** |
0.06 |
Yes |
|
H9: PEOU |
0.32** |
0.04 |
Yes |
|
H10: PU |
0.09 |
0.13 |
No |
|
H11: ATU |
0.29* |
0.13 |
Yes |
|
H12: PU |
0.49*** |
0.13 |
Yes |
PUNote: ***p < 0.001,**p < 0.01, *p < 0.05
6.1 Discussion
The growing popularity of RS in educational institutions has increased awareness of its use in the classroom, which is centred on the concept of personalisation. Additionally, recommender systems are offered to maximise a learner's learning experience because course selection has a direct impact on a learner's performance.In this connection, one of the first measures has been to look into students' adoption of RS. Current research, on the other hand, fails to uncover qualities that may influence students' adoption of RS, as well as factors that may be influenced by their usage of RS. Therefore, the proposed comprehensive model was enhanced to bridge the gap.This study looked into the elements that influenced students' acceptance of and use of a course recommendation system in higher education. The proposed study methodology was built around the TAM characteristics of perceived utility, perceived ease-of-use, and behavioural intention. External elements were added to TAM to forecast the fundamental TAM constructs of Availability, Experience, and Relevance. Structural equation modelling was used to investigate the interactions between these constructs.
Perceived Usefulness and Perceived Ease of Use are both significant predictors of Attitude towards Using and Behavioural Intention, according to the findings.The findings are consistent with those of prior TAM research (Liu, Hone, &Tarhini, 2014; Tsai &Weng, 2015).Furthermore, perceived usefulness is found to be a stronger interpreter of behavioural intent than perceived ease of use, which is in line with earlier studies.(Kang & Shin, 2015; Al-Azawei, 2017).When it comes to the intention to utilise RS in course selection, learners are less hesitant to use it and are more likely to focus on its use and benefits.When it comes to the intention to use RS in course selection, learners who feel free to utilise RS are more likely to focus on the utility and benefits of doing so. It may be deduced that, despite the fact that a vast majority of students still lack access to RS, students' trust in the system remains a concern.Another evidence to understand student varied impressions of using RS could be the surprisingly unsupported hypothesised association between Experience with Attitude toward utilising and Relevance with Perceived Usefulness.
In addition to the core TAM elements, external factors such as availability, experience, and relevance are introduced to improve the model's explanatory power. Experience has been shown to be a key predictor of Behavioural Intention to Use, and its expansion prove more beneficial perceptivenessand this outcome is consistent with previous study (Wang, Yan, &Su, 2017).Learners who have more familiarity with RS may be more familiar with the types of courses offered by the system. By detecting appropriate courses based on learner inquiries, RS benefits can be maximised, as it is fully semantic and ideal for the learners' future development. As a result, as the learner's experience with RS grows, so does his or her productivity and quality of learning. This conclusion supports Kenny and Way's(2001) assertion that "Once learners have experience of what RS can and cannot do, learners can reflect on the function of such systems in the recommendation process in a more realistic and less defensive way". Another piece of evidence that could help explain learner perceptions of utilising RS is the surprisingly projected link between relevance and perceived ease of use.Despite the fact that the uncorroborated hypothesis contradicts Venkatesh er al; Technology acceptance model, it is consistent amid previous research (Leong et al., 2018). One probable elucidation is that RS is deemed simple to use, which does not necessarily imply that learners find it useful. As a result, greater effort should be put into understanding the antecedents of learner’s perceptions of usefulness.In addition to the analysis of unique TAM structures, RS availability is integrated as an external component to increase the model's instructional power. The availability of a RS is seen to be a major predictor of perceived usefulness and attitude toward using it. Increasing experience, availability, and relevance could lead to higher judgments of usefulness for all three factors.Besides, the subsequentresult of using RS is confirmed. Learners' attitudes toward utilising RS can be influenced by perceived usefulness, perceived ease of use, availability, and relevance. Because RS is primarily connected with ease of use, easy accessibility and quick reaction are not difficult to comprehend. If a learner finds RS to be simple to use, they will be driven to use it to improve their subject knowledge and will strongly encourage others to use RS while selecting courses and performing other learning tasks. Furthermore, experience is discovered to be a strong predictor of RS Behavioural Intention. The outcomes of this study revealed important elements that influence theadoption of RS and the advantages of doing so.
6.2. Implications
In spite of learners' increased attention in tools that propose courses for their study, studies investigating the other external elements for RS's adoption behaviors in suggesting courses at the educational level, particularly at the research and master’s level, are still lacking.The current research could help with both theoretical and practical applications of RS research. The expanded quasi-circular model with Availability, Experience, and Relevance, in theory, aids in identifying the variables that influence RS adoption and those that are influenced by employing RS. Furthermore, being a novel recommendation technology in education, RS's adoption research demands a robust and thorough model. As a result, this research could help enhance TAM theory in RS research.In practice, the findings of this study may offer advice to faculty, researchers, policymakers, and RS developers. Despite the fact that TAM has been widely used in e-learning studies, more information about content-specific and contextual usage is required (Teo, Huang, & Hoi, 2018).By giving enormous and overwhelming influence of learners perceived utility about their intent to use RS, RS developers must be conscious of technological limits that affect recommendation quality, and for faculties to be aware of the non-technological antecedents of Perceived Usefulness, Attitude towards using and behavioural intention.Instructing learners to increase their use of RS and documenting their views, for example, could be a beneficial strategy to improve recommendation algorithms, given that experience has been shown to be a substantial predictor of behavioural intention.Furthermore, Perceived Usefulness and Perceived Ease of Use are shown to have a significant impact on students' attitudes toward RS, whereas Experience may have an impact on their behavioural intentions. This research suggests that the conceptions have a cyclical influence relationship, which could lead to pedagogical recommendations for faculty.Using RS as an instance, faculties might use it as an assisting tool for improvinglearner’s outcomes. Moreover, faults and inconsistencies in RS outputs might be exploited to improve systems in order to make them more competent. Finally, the outcomes of this study may help policymakers better comprehend RS and become more conscious of how to incorporate it into educational sites and portals.
6.3 Limitations
However, there are certain limitations to this study that should be addressed in upcoming research studies. Primarily, the studied' sample sizes are smaller that may restrict the findings to be generalized.In future, the size of the sample taken should be bigger so that it will be examined to improve expounding power the models.Second, the majority of the study's participants are beginner researcher scholars. According to Teo et al(2014),situation-based and population external constructs may contribute to various TAM outcomes. Learners with various domain and knowledge backgrounds about RS will be involved in subsequent study to further validate the findings. Finally, this study's conclusions are based on self-reported and cross-sectional data. Students' attitudes toward using RS may evolve over time. To capture the desire to utilise RS, a long term study with a differentsources of data (e.g. Qualitative data: interviews and observations) would be examined.
To learn about the aspects that goes into successfully integration of recommendation systems in educational domain. The goal of this research is to figure out whether scholars will accept the recommendations suggested by the RS and what are the factors that influence a student or scholar to embrace new technology. The goal is to learn more about scholar or students' perceptions of the usefulness of recommendation system and the factors that impact their decision to integrate and embrace technology into their learning system.In this study it was seen that the scholars found the RS very useful in terms of choosing a course suggestions as they very much relevant and would be helpful for their learning needs. One more thing was observed by the scholars who have already completed their course work, that the courses they have as chosen manually was very much hectic, time consuming and not appropriate as per their learning needs. The study, which includes a customised original TAM, aims to assess research experts' and students' willingness to use a recommender system. The TAM's basic constructs were modified to validate the relationship between perceived usefulness, perceived ease of use, attitude toward usage, and the impact on behavioural intention to use in general. No unexpected findings were reported in earlier constructs; hence the current study verifies past findings and empirical evidence based on the technological acceptance paradigm. It also successfully proves TAM's applicability to recommender systems. The findings of this study indicated that TAM is a reliable model for acceptance that may be used to predict behavioural intentions toward RS use before it is implemented.Furthermore, the study model has been validated in the setting of higher education with the highest level of the group. Furthermore, this research may aid institutions and E-learning software manufacturers in their efforts to implement such features, thereby assisting in the promotion of E-learning in poor nations.
Funding and Conflicts of Interest Statement:
Manuscript title: “Integrating External Factors and Technology Acceptance Model to Understand Scholar Intention and Use of Recommendation System for Course Selection.”The authors whose names are listed immediately below certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.
Author names: Dr. ZameerGulzar, ,Dr.FatimaAmerJidAlmahri, Afrahfathima
Please address all correspondence concerning this manuscript to me at [email protected]
- Abidi, M. S. H. D., Khan, M. N. “An analysis into UTAUT based research models for IT implementation”. Studies in Indian Place Names, 40(71), 2011-2027. (2020).
- Adomavicius, G., &Tuzhilin, A. (2005). Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions. IEEE Transactions on Knowledge and Data Engineering, (6), 734–749. https://doi.org/10.1109/TKDE.2005.99.
- Ajzen, I., and Fishbein,M.. (1980). Understanding Attitudes and Predicting Social Behavior.Englewood Cliffs, NJ: Prentice-Hall.
- Aguilar, J., Valdiviezo-Díaz, P., &Riofrio, G. (2017).A general framework for intelligent recommender systems. Applied Computing and Informatics, 13(2), 147–160. https://doi.org/10.1016/j.aci.2016.08.002.
- Ahmed, T.S., Kamal, M.B., Nik Suryani, A. and Tunku, B.T.A. ‘Investigating students’ attitude and intention to use social software in higher institution of learning in Malaysia’, Multicultural Education and Technology Journal, Vol. 5, No. 3, 2011. pp.194–208
- Alpert, S.R., J. Karat, C.M. Brodie Karat., and Vergo J.C.(2003),“User attitudes regarding a user -adaptive e-Commerce web site. ”User Modeling and User Adapted Interaction, Vol.13, No. 4,pp. 373-396.
- Al-Azawei, A. (2017). Investigating the effect of learning styles in a blended e-learning system: An extension of the technology acceptance model (TAM). AustralasianJournal of Educational Technology, 33(2), 1–23. https://doi.org/10.14742/ajet.2741.
- Anderson, D. L. (2004).The Textualizing Functions of Writing for Organizational Change. Journal of Business and Technical Communication, 18(2), 141–164. https://doi.org/10.1177/1050651903260800
- Anderson, J. C., &Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two step approach. Psychological Bulletin, 49(2),411–423. https://doi.org/10.1037//0033-2909.103.3.411.
- Armentano, M.G., Abalde, R., Schiaffino,S., Amandi, A.(2014). User acceptance of recommender systems: Influence of the preference elicitation algorithm. In 9th International Workshop on Semantic and Social Media Adaptation and Personalization (SMAP), pp. 72–76.
- Asosheh, A., Bagherpour, S., and Yahyapour, N.(2008). Extended acceptance models for recommender system adaption, case of retail and banking service in Iran. WSEAS Transactions on Business and Economics, Vol. 5(5), pp. 189–200.
- Baier, D., and Stüber, E. (2010). Acceptance of recommendations to buy in online retailing.Journal of Retailing and Consumer Services, Vol. 17(3), pp. 173–180.
- Bentler, P. M., & Chou, C. P. (1987).Practical issues in structural modeling. Sociological Methods & Research, 16(1), 78–117. https://doi.org/10.1177/0049124187016001004
- Burdea, G.C., &Coiffet, P. (2003). Virtual reality technology, (2nd ed.), New York: John Wiley & Sons.
- Byrne, B. M. (2013). Structural equation modeling with AMOS: Basic concepts, applications, and programming. New York: Routledge.
- Cacciamani, S., Villani, D., Bonanomi, A., Carissoli, C., Olivari, G. M., Morganti, L., et al. (2018). Factors affecting students' acceptance of tablet PCs: A study in Italianhigh schools. Journal of Research on Technology in Education, 50(2), 120–133. https://doi.org/10.1080/15391523.2017.1409672
- Cheung, R., & Vogel, D. (2013).Predicting user acceptance of collaborative technologies: An extension of the technology acceptance model for e-learning. Computers &Education, 63, 160–175. https://doi.org/10.1016/j.compedu.2012.12.003.
- Chen, H.C., Hsu, C.C., Chang, C.H., and. Huang, Y.M. (2012). Applying the technology acceptance model to evaluate the learning companion recommendation system on Facebook.In IEEE Fourth International Conference on Technology for Education (T4E), pp. 160–163.
- Chughtai, M. W., Selamat, A., Ghani, I., & Jung, J. J. (2014).Retracted: E-Learning recommender systems based on goal-based hybrid filtering.International Journal of Distributed Sensor Networks, 10(7), 912130. https://doi.org/10.1155/2014/912130.
- Chuttur,M. (2009). Overview of the technology acceptance model: Origins, developments and future directions. Working Papers on Information Systems, Vol. 9(37), pp. 1–22, 2009.
- Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. https://doi.org/10.1037/0033-2909.112.1.155.
- Dascalu, M. I., Bodea, C. N., Mihailescu, M. N., Tanase, E. A., &Ordoñez de Pablos, P. (2016). Educational recommender systems and their application in lifelong learning. Behaviour & Information Technology, 35(4), 290–297.
- Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology.MIS quarterly, 13(3), 319-340.
- Davis, F. D. (1993). User acceptance of information technology: System characteristics, user perceptions and behavioural impacts. International Journal of Man-Machine Studies, Vol.38, pp.475-487.
- Davis, F., Bagozzi, R., & Warshaw, P. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982-1003.
- Esteban-Millat, I., Martínez-López, F. J., Pujol-Jover, M., Gázquez-Abad, J. C., &Alegret, A. (2018).An extension of the technology acceptance model for onlinelearning environments. Interactive Learning Environments, 26(7), 895–910. https://doi.org/10.1080/10494820.2017.1421560
- Farrell, T., &Rushby, N. (2016). Assessment and learning technologies: An overview. British Journal of Educational Technology, 47(1), pp.106–120.
- Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research. 1981; 18(1):39–47.
- Gil A B, Garciapenalvo F J. (2008) “Learner Course Recommendation in e-Learning Based on Swarm Intelligence”Journal of Universal Computer Science: 2737-2755.
- Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998).Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall.
- Hair, J., Black, W0.,Babin, B., Anderson, R., and Tatham, R. (2006). Multivariate Data Analysis: Pearson Education," New Jersey: Hoboken.
- Hazra, I., Quang, H., Ting-Wen, C., Kinshuk, Sabine, G. (2014) .A Framework to Provide Personalization in Learning Management Systems through a Recommender System Approach.Springer International PublishingSwitzerland, pp. 271–280.
- Hsiu-Mei Huang and Shu-Sheng Liaw(2018) International Review of Research in Open and Distributed Learning Volume 19, Number 1 .An Analysis of Learners’ Intentions Toward Virtual Reality Learning Based on Constructivist and Technology Acceptance Approaches.pp. 91-115
- Hubert, M., Blut, M., Brock, C., Backhaus, C., &Eberhardt, T. (2017). Acceptance of Smartphone-Based Mobile Shopping: MobileBenefits, Customer Characteristics, Perceived Risks, and the Impact of Application Context. Psychology and Marketing, 34(2),175–194. https://doi.org/10.1002/mar.20982
- Hughes, C., &Leekam, S. (2004). What are the Links Between Theory of Mind and Social Relations? Review, Reflections and New Directions for Studies of Typical and Atypical Development. Social Development, 13(4), 590–619. doi:10.1111/j.1467-9507.2004.00285.x
- Hu, R., and Pu, P. (2009). Acceptance issues of personality based recommender systems, In Proceedings of ACM RecSys’09. New York, NY, USA: ACM, pp. 221–224.
- Hung, Y.H., Hu, P.C., and Lee, W.T.(2013). Improving the design and adoption of travel websites: A user experience study on travel information recommender systems, In 5th IASDR International Conference, Japan.
- Jemain, A. A., Al-Omari, A., and Ibrahim, K. (2007). Multistage median ranked set sampling for estimating the population median. Journal of Mathematics and Statistics, Vol. 3, pp. 58-64.
- JinjiaoLina,b,, HaitaoPuc, Yibin Lib Jian Lianc, 2018. “Intelligent Recommendation System for Course Selection in Smart Education”, Procedia Computer Science 129 (2018) 449–453, DOI: 10.1109/ICDM.2011.134
- Kang, M., & Shin, W. (2015).An empirical investigation of student acceptance of synchronous e-learning in an online university. Journal of Educational ComputingResearch, 52(4), 475–495. https://doi.org/10.1177/0735633115571921.
- Kenny, D., & Way, A. (2001). Teaching machine translation and translation technology: A contrastive study. Proceedings of the machine translation summit VII, teachingMT workshop (pp. 13–17). Spain: Santiago de Compostela.
- Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York: The Guilford Press.
- Leong, L. W., Ibrahim, O., Dalvi-Esfahani, M., Shahbazi, H., &Nilashi, M. (2018). The moderating effect of experience on the intention to adopt mobile social networksites for pedagogical purposes: An extension of the technology acceptance model. Education and Information Technologies, 23, 2477–2498. https://doi.org/10. 1007/s10639-018-9726-2.
- Lee, H.Y., Ahn, H., and Han, I. (2007). VCR: Virtual community recommender using the technology acceptance model and the user’s needs type. Expert Systems with Applications, Vol. 33(4), pp. 984–995.
- Liébana-Cabanillas, F., Muñoz-Leiva, F., Sánchez-Fernández, J., &Viedma-del Jesús, M. I. (2016). The moderating effect of user experience on satisfaction withelectronic banking: Empirical evidence from the Spanish case. Information Systems and e-Business Management, 14, 141–165. https://doi.org/10.1007/s10257-015-0277-4.
- Legris, P., Ingham, J., &Collerette, P. (2003). Why do people use information technology? A critical review of the technology acceptance model. Information &Management, 40(3), 191–204. https://doi.org/10.1016/S0378-7206(01)00143-4
- Lu, J., Wu, D., Mao, M., Wang, W., & Zhang, G. (2015). Recommender system application developments: A survey. Decision Support Systems, 74, 12–32. https://doi.org/10.1016/j.dss.2015.03.008.
- Marangunić., N, and Granić., A. (2015). Technology acceptance model: A literature review from 1986 to 2013. Universal Access in the Information Society, 14(1), pp.81–95.
- Nilashi, M., Ibrahim, O., &Bagherifard, K. (2018). A recommender system based on collaborative filtering using ontology and dimensionality reduction techniques. Expert Systems with Applications, 92, 507–520. https://doi.org/10.1016/j.eswa.2017.09.058.
- Núñez-Valdéz, E. R., Lovelle, J. M. C., Martínez, O. S., García-Díaz, V., De Pablos, P. O., &Marín, C. E. M. (2012). Implicit feedback techniques on recommender systems applied to electronic books. Computers in Human Behavior, 28(4), 1186–1193. https://doi.org/10.1016/j.chb.2012.02.001.
- O’Leary, Z. (2017) The Essential Guide to Doing Your Research Project. Sage, London.
- Potkonjak, V., Gardner, M., Callaghan, V., Mattila, P., Guetl, C., Petrovic, V. M., et al. (2016). Virtual laboratories for education in science, technology, and engineering:A review. Computers & Education, 95, 309–327. https://doi.org/10.1016/j.compedu.2016.02.002
- Pu, P., Chen, L., and Hu, R. (2011). A user-centric evaluation framework for recommender systems.In Proceedings of the fifth ACM conference on Recommender systems. ACM, pp. 157–164.
- Pu, P., Chen, L., and Hu, R. (2012).Evaluating recommender systems from the user’s perspective: survey of the state of the art.User Modeling and User-Adapted Interaction, Vol. 22 (4-5), pp. 317–355.
- Rauniar., R, Rawski., Yang., G.J, &Johnson.,B. (2014). Technology acceptance model (TAM) and social media usage: An empirical study on Facebook. Journal of Enterprise Information Management, 27(1), pp.6–30.
- Renaud, K., & Van Biljon, J. (2008). Predicting technology acceptanceand adoption by the elderly: A qualitative study. InProceedings of the 2008 Annual Research Conference of theSouth African Institute of Computer Scientists and InformationTechnologists on IT Research in Developing Countries: Ridingthe Wave of Technology (pp. 210–219). New York, NY: ACM.doi:10.1145/1456659.1456684.
- Ricci, F., Rokach, L., Shapira, B. (2011).Introduction to Recommender Systems Handbook.Springer, ch.1, pp. 1–35.
- Ronnie, H.S.,Christopher, C.D., Eugenia M.W. (2011). Analysis of the technology acceptance model in examining students' behavioural intention to use an e-portfolio system.Australasian Journal of Educational Technology, Vol. 27, pp. 600-618.
- Sánchez, R. A., and Hueros, A. D. (2010). Motivational factors that influence the acceptance of Moodle using TAM. Computers in Human Behavior, Vol. 26, pp. 1632-1640.
- Sekaran U. (2006). Research methods for business: A skill building approach: New York, USA: John Wiley & Sons. 2006.
- Selim, H.M. (2003). An empirical investigation of student acceptance of course websites.Computers & Education, 40(4), 343-360.
- Shih, H.P. (2004). Extended technology acceptance model of Internet utilization behaviour.Information & Management, Vol. 41, pp. 719-729, 2004.
- Stoel, L., &Hye Lee, K. (2003).Modeling the effect of experience on student acceptance of web-based courseware. Internet Research, 13, 364–374. https://doi.org/10. 1108/10662240310501649.
- Su, P., Wang, L., & Yan, J. (2017). How users' Internet experience affects the adoption of mobile payment: A mediation model. Technology Analysis & StrategicManagement, 30(2), 186–197. https://doi.org/10.1080/09537325.2017.1297788.
- Swearingen, K., and Sinha, R. (2001). Beyond algorithms: An HCI perspective on recommender systems,” In ACM SIGIR Workshop on Recommender Systems, Vol. 13, pp. 393–408.
- Tarhini, A., Hone, K., & Liu, X. (2014).Measuring the moderating effect of gender and age on e-learningacceptance in England: A structural equation modelingapproach for an extended technology acceptance model. Educational Computing Research, 51(2), 163–184. https://doi.org/10.2190/EC.51.2.b.
- Taylor, S., and Todd, P. (1995).Assessing IT usage: The role of prior experience.MIS quarterly, Vol. 19, pp. 561-570.
- Teo, T., Huang, F., & Hoi, C. K. W. (2018).Explicating the influences that explain intention to use technology among English teachers in China. Interactive LearningEnvironments, 26(4), 460–475. https://doi.org/10.1080/10494820.2017.1341940
- Venkatesh, V., & Davis, F. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186–204.
- Venkatesh, V., Morris, M., Davis, G., & Davis, F. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27, 425–478.
- Weng, C., & Tsai, C. (2015).Social support as a neglected e-learning motivator affecting trainee's decisions of continuous intentions of usage. Australasian Journal of Educational Technology, 31(2), 177–192. https://doi.org/10.14742/ajet.1311.
- Wong, G. (2016). The behavioral intentions of Hong Kong primary teachers in adopting educational psychology. Educational Technology Research & Development, 64, 313–338. https://doi.org/10.1007/s11423-016-9426-9.
- Yanxia Yang, Xiangling Wang (2019). “Modeling the intention to use machine translation for studenttranslators: An extension of Technology Acceptance Model”, Computers & EducationComputers &Educationol(133), pp.116-126. https://doi.org/10.1016/j.compedu.2019.01.015
- Zhu, Z., Yu, M. &Riezebos, P.A research framework of smart education.Smart Learning Environments.3, 4 (2016).https://doi.org/10.1186/s40561-016-0026-2
- Zouhair, J. (2012). Surveying Learners’ Attitudes Toward a Saudi E-learning System.International Journal of Information and Electronics Engineering, Vol. 12, September 2012.
- ZuhalHusseina. (2017),” Leading to Intention:The Role of Attitude in Relation to Technology Acceptance Model in E-Learning”, Procedia Computer Science, Vol 105,159 – 164, doi: 10.1016/j.procs.2017.01.196
No competing interests reported.