نوع مقاله : مقاله پژوهشی (کاربردی)
نویسندگان
1 استادیار، گروه مدیریت دانش، دانشکده علوم اجتماعی، دانشگاه فرماندهی و ستاد آجا.
2 استادیار، گروه ریاضی، دانشکده علوم پایه، دانشگاه آزاد واحد الکترونیکی.
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Today, with the increase in the use of the internet, people have turned to the internet to buy their products or to learn about various topics. There are a large number of virtual pages where users post their opinions on various topics. A large amount of data exists which extracting useful information from is a costly and time-consuming task. Opinion mining is the process of intelligent analysis of the sentiments of users who have expressed their opinions in relation to a specific topic with the capability of extract them. The machine learning method is one of the most optimal and efficient methods for extracting knowledge from users' opinions on the offered products. In these methods, the training data of a system is given to classify user opinions. One of the most important classification steps is data reduction. By using the new feature combination method, the set of extracted features can be reduced to a greater extent than the feature selection method, which leads to a subset of useful information with a much smaller volume and higher recognition power. In this research, particle group optimization algorithm was used to optimize the combination of features. To evaluate the proposed method, MATLAB software was used to evaluate the proposed method, and experiments were conducted on four data sets. The results of the research showed that the use of the feature combination method increased the efficiency of classification and reduced the effect of this increase in the decrease of the efficiency of the classifier.
کلیدواژهها [English]