In today’s scenario where a vast amount of text, images and documents are sent and received mostly of which the data in unstructured textual form. We observe that every minute of the day a large amount of data is stored. This huge amount of data needs to be process which is impossible for human being to do it. The purpose of this comparative analysis is to reduce the classification errors using various algorithms by estimating the distribution of class by using vectors. Nowadays, to acquire meaningful, useful data from the vast variety of textual and documented data present it is necessary and viable to build techniques and algorithm better than before. Hence, in order to get interesting and needful information we use model classification, evolution and regression. The data is collected from records or created or taken from the organization to create the database then the pre-processing of data is done so that various classification algorithms can be applied and can be compared on the basis of precision, efficiency, accuracy, ROC curve and missing values and different results can be predicted to solve the problems.