Opinion mining on film critique using a combination of support vector machine and particle swarm optimization
Saman Siadati *, M. Jafar Tarokh **
The online social media is actually an area for exchanging views and opinions at a moment's notice, especially when people share and ultimately share content. One of the fastest and easiest way to send messages on the social media is Twitter. Messages on Twitter include comments on specific topics such as movies, books, goods, politics and other topics. Based on this, the research is attempting to utilize Twitter's messaging system using Opinion Mining or Sentiment Analysis. This discussion deals with the use of natural language processing, computational linguistics, or text mining to identify or classify whether the film in question is a good film, based on the comment sent. Support Vector Machine is based on supervisory learning methods that analyzes the data and identifies the patterns used for categorization. The main concern of this study is the classification of binary classified into two positive and negative classes. The positive class represents positive comments, and in contrast to the negative class, it represents a negative or negative message about certain films. This justification is based on the level of precision of the class of vector support machine with the validation processes. In order to select the best parameter for solving the dual optimization problem, particle swarm optimization algorithm has been used. The result shows an improvement in the accuracy level to an acceptable level.
Opinion Mining, Sentiment Analysis, Support Vector Machine, Particle Swarm Optimization algorithm
*,** Strategic Intelligence Research Laboratory, K.N.T. university of technology, Tehran, Iran