Investigation of the Effects of Silage Type, Silage Consumption, Birth Type and Birth Weight on Live Weight in Kıvırcık Lambs With MARS and Bagging MARS Algorithms.

Investigation of the effects of silage type, silage consumption, birth type and birth weight on live weight in Kıvırcık lambs with M ARS and Bagging MARS algorithms Abstract: This study was carried out to determine the effect of silage type, silage consumption, 22 birth type (single or twin) and birth weight on live weight at the end of fattening in Kıvırcık 23 lambs. In the experiment, 40 male Kıvırcık lambs aged 2.5-3 months were used and the animals 24 were fattened for 56 days. During the fattening period, the lambs fed with 5 different types of 25 silage (100% sunflower silage, 75% sunflower + 25% corn silage, 50% sunflower + 50% corn 26 silage, 25% sunflower + 75% corn silage, 100% corn silage) pure and mixed in different 27 proportions and concentrate feed. Data on fattening results were analyzed with MARS and 28 Bagging MARS algorithms. The main objective of this research is to predict live weight of 29 lambs using Multivariate Adaptive Regression Splines (MARS) and Bagging MARS 30 algorithms as a nonparametric regression technique. Live weight value was modeled based on 31 factors such as birth type, birth weight, silage type and silage consumption. Correlation 32 coefficient (r), determination coefficient (R 2 ), Adjust R 2 , Root-mean-square error (RMSE), 33 standard deviation ratio (SD ratio), mean absolute percentage error (MAPE), mean absolute 34 deviation (MAD), and Akaike Information Criteria (AIC) values of MARS algorithm predicting 35 live weight were as follows: 0.9986, 0.997, 0.977, 0.142, 0.052, 0.2389, 0.086 and -88 36 respectively. Like statistics for Bagging MARS algorithm were 0.754, 0.556, 0.453, 1.8, 0.666, 37 3.96, 1.47 and 115 respectively. It was observed that MARS and Bagging MARS algorithms 38 have revealed correct results according to goodness of fit statistics. However, it has been 39 revealed that MARS algorithm gives better results in live weight modeling.


Investigation of the effects of silage type, silage consumption, birth type and birth weight
Approximately 70% of the expenses of the enterprises engaged in animal production are 44 roughage and intensive feed costs (Kara and Eroğlu, 2018). This is very important in terms of 45 showing how effective and decisive the feed is in the development of livestock. Today, where 46 the demand for animal products is increasing, more and more roughage and concentrate feed production is needed for more animal food production.
In order to obtain high efficiency from animals, it is necessary to meet the nutrient needs in a 49 balanced and sufficient level, and for this purpose, it is necessary to use quality roughage and 50 concentrate feed sources. Roughage is generally divided into two groups as dry and watery 51 roughage. Dry roughage, hay, straw and products with a crude cellulose content of 18% or 52 higher, roughage consists of green fodder plants such as alfalfa, sainfoin, vetch, silage, roots 53 and tubers. 54 One of the main problems of animal husbandry is the difficulties in obtaining good quality, It is possible to benefit from sunflower as an important forage plant, thanks to its ability to be 82 silage in a shorter time than corn, its tolerance to high and low temperatures, and its high 83 adaptability to various soil conditions (Yıldız, 2017).

84
Although silage is one of the most important roughage sources used in the feeding of sheep and 85 goats as well as cattle in countries with developed livestock, silage production and use are still 86 insufficient in some countries. Especially, the use of silage is very low in small ruminant.

87
However, it has been reported that silage feed has started to be used in the rations of small 88 ruminant in recent years (Öztürk, 2000).

89
In this study, it was aimed to investigate the effects of different silage type, silage consumption, here is the number of basis functions, and is the penalizing parameter. The optimal value 151 of usually falls in the range of 2 ≤ ≤ 4, and generally = 3 is used (Friedman, 1991).  As seen in Table 1, 13 of the 40 Kıvırcık lambs were born as singles and 27 of them were twins.

203
The mean birth weight of the lambs was 4.64 kg in single and 4.03 kg in twins. While single 204 lambs consumed an average of 917 g of silage per day, twin lambs consumed 932 g of silage.

205
The average live weight at the end of fattening is 37.5 kg in single lambs and 36.6 kg in twin 206 lambs. The values belonging to single and twin lambs is presented in Figure 1. In   The relative importance of the independent variables is presented in Table 5.
268 269 Table 5. Relative importance of model independent variables.

270
As seen in The distribution graphs of observed predicted values of LW was indicated in Figure 3. The plot between the predicted and observed LW values is showed in Figure 4