การสกัดคุณลักษณะข้อความที่มีประสิทธิภาพเพื่อการจำแนกขั้วความคิดเห็น

Abstract

Recently, social media users can comment with texts to describe their opinions. These texts can be analyzed to classify them into positive and negative directions. Before creating classifier, the feature vectors for representing the texts must be prepared firstly. Generally, texts are represented by vectors of weights or frequencies of terms that appear in the text. The number of dimensions of vector is equal to the number of terms in the dictionary derived from the possible words in all texts. The large amount of words in dictionary leads to the high dimensional vector for representing text and bring about the long processing time to training and testing the text classification models. This thesis proposed two methods for representing texts including V4D and V8D which are the low-dimensional vectors. The set of positive and negative words were considered to create the vectors. In addition, the feature vectors were derived by using the words of negation which have the significant meanings in a classification of text opinions. In this thesis, four classification techniques including k-Nearest Neighbors, Naive Bayes, Artificial Neural Networks and Support Vector Machine were studies to classify the opinion texts. By experimenting on eight data sets with various domains, the proposed vectors, including V4D and V8D, were compared with the traditional vectors, including TF and TF-IDF in the view of the performances when they were applied to the classification problem. The experimental results show that the proposed vectors for representing text can improve the performance of opinion text classification and provide the best efficiency in the terms of used space and processing time.