Comparison of Machine Learning Classification Algorithms for Purchasing Forecast

Abstract

With the development of computer technologies and invention of internet, many concepts have entered our lives. With the starting of wide usage of globalized internet network, concept of machine learning has emerged in time for smarter management of data flow in big dimensions. In line with technological developments, all activities began to be carried to digital environment and as a result of this, concept of e-commerce has entered our lives. E-commerce is one of the areas where machine learning is used most widely. By examining product purchasing situations in accordance with data available at the enterprises, various researches have been made for selection of most appropriate model in order to predict future data. In the study it was mentioned about concepts of e-commerce and machine learning and by applying Logistic Regression, Naïve Bayes and Support Vector Machines being machine learning classification algorithms, it has been aimed to determine the model having best accuracy ratio.

Classification Algorithm for Heterogeneous Network Data Streams Based on Big Data Active Learning

“…After converting the textual data into digital features, the training time was reduced with an accuracy of 87.08%, saving 87.97% of the training time [15]. Zdemir et and Turani proposed a machine learning approach applicable to the e-commerce domain, applying Logistic Regression, Parsimonious Bayes, and Support Vector Machines as data classification algorithm, aiming to identify the model with the best accuracy [16]. Qiu et al proposed a data stream classification model based on distributed processing in order to solve the problem of real-time detection of grid equipment anomalies.…”