International Journal of Computer Applications (0975 – 8887) International Conference on Cognitive Knowledge Engineering 2016 8 Mining Movie Intention using Bayes and Maximum Entropy Classifiers Varsha D. Jadhav Assistant Professor P.E.S. College of Engineering Aurangabad, (MS) India Sachin N. Deshmukh Professor Department of CS/IT, Dr. Babasaheb Ambedkar Marathwada University Aurangabad, (MS) India ABSTRACT Sentiment analysis is becoming one of the most thoughtful research areas for prediction and classification. This paper analyzes and predicts the result for movie reviews. Machine learning techniques Bayes and Maximum entropy for classifying text messages. Movie comments from twitter are retrieved. The two classifiers are analyzed for Hindi movies ‘Sultan’ and ‘Madaari.’ Tweets before and after the release of the movie are retrieved. Accuracy is evaluated to compare the Bayes and Maximum entropy methods. R technology is used for the movie review analysis. General Terms Intention Mining, machine learning, normalization, Twitter API. Keywords Bayes, Maximum Entropy, polarity, emotions, mean absolute error. 1. INTRODUCTION Social networks today contains enormous amount of text data, which is growing everyday. Intention mining aims to cover the attitude of the author on a particular topic from text data. It is natural language processing and machine learning techniques reveal the attitude. In the recent years it has gained popularity due to its immediate application in business, customer feedback from product reviews, spots reviews and assisting in election campaigns. Movie reviews are a important way to analyze the performance of a movie. Text movie reviews tells us the strong and weak points of the movie which tells us whether the movie in general meets the expectations of the reviewer. Using intention mining, we can find the state of mind of the reviewer and understand the polarity and emotions. Polarity is positive, negative and neutral. Emotions are anger, disgust, fear, joy, sadness, surprise, and unknown. In this paper we use intention mining on a set of movie reviews extracted from twitter and try to understand the overall reaction about the movie. Whether people liked or they disliked it. We analyzed the movie reviews using machine learning methods Bayes and maximum entropy. We compare the two methods for accuracy. 2. RELATED WORK Minqing Hu and Bing Liu [1], studied the problem of generating feature-based summaries of customer reviews of products sold online. Here, features broadly mean product features (or attributes) and functions. Given a set of customer reviews of a particular product, the task involves three subtasks: (1) identifying features of the product that customers have expressed their opinions on (called product features); (2) for each feature, identifying review sentences that give positive or negative opinions; and producing a summary using the discovered information. Varsha D. Jadhav and S.N. Deshmukh [2] presented the Bayes model to predict the intentions of the people who tweet on twitter about a specific topic. They predicted the cricket match result which serves as strategic guidance to the captains so as to improve the performance of the team. Kamal Nigam et.al [3] proposes the use of maximum entropy techniques for text classification. In their text classification scenario, maximum entropy estimates the conditional distribution of the class label given a document. A document is represented by a set of word count features. The labeled training data is used to estimate the expected value of these word counts on a class-by-class basis. Improved iterative scaling finds a text classifier of an exponential form that is consistent with the constraints from the labeled data. Kuat Yessenov, et.al [4], presented an empirical study of efficacy of machine learning techniques in classifying text messages by semantic meaning. They used movie review comments from popular social network Digg as the data set and classify text by subjectivity/objectivity and negative/positive attitude. Changlin Ma, et.al [5], proposed a novel topic and sentiment unification maximum entropy LDA model in this paper for fine-grained opinion mining of online reviews. Oaindrila Das, et.al [6] presented a novel approach for classification of online movie reviews using parts of speech and machine learning algorithms. Borislav Kapukaranov and Preslav Nakov[7] presented experiments in predicting fine- rained stars, including halves, for Bulgarian movie reviews. This is a challenging task, that can be seen as (a) multi-way classification, i.e., choosing one out of eleven classes, (b) regression, i.e., predicting a real number, or (c) something in between, namely ordinal regression, i.e., predicting eleven values, but taking ordering into account, e.g., predicting 4 when the actual value is 3.5 would be better than predicting 1. 3. METHODOLOGY Fig. 1 shows the framework for intention mining system.