زهرا احمدیان

  Due to the fact that the growth of data production in today's human life has already accelerated, therefore, different methods have been expressed in this direction in order to study data, extract valuable insights and information. Data is considered one of the biggest assets of any company, institution, researchers, etc., whose analysis helps the process of making important decisions and forecasting. One of the basic methods used for data analysis is data classification, which is a subset of supervised learning. The meaning of classification is that the data is labeled under the eyes of news people, and try to use the set of training relationships between the input special attributes with one or more special attributes to discover and build a model for predicting new labels. It is the purpose of it. A model is created to sort the data into predefined classifications so that the number of classifications is determined in advance. Different methods and algorithms in the field of classical classification have been created by different researchers since the past. In this thesis, we intend to compare the boosting classification method using CatBoost algorithm with classical classification methods such as support vector machine, logistic regression,$k-$ nearest neighbor with Bayesian methods such as Bayesian support vector machine,combinition of $k-$ nearest neighbor with Naive Bayes and Bayesian logistic regression.