THE FIFTH INTERNATIONAL CONFERENCE ON ICT IN OUR LIVES, 19-21 DECEMBER 2015 102 OpinionMining Based on Text Mining in Social Networking Framework Hatem Abdel Kader Department of Information Systems Menoufyia University Faculty of Computers and Information Shebien El Koum, Egypt hatem6803@yahoo.com Abd El-Fatah Hegazy Department of Information Systems Arab Academy for Science and Technology and Maritime Transport Cairo, Egypt abdheg@yahoo.com Fainan Nagy El sisiDepartment of Information Systems Arab Academy for Science and Technology and Maritime Transport Cairo,Egypt fainan_nagy1@yahoo.com Abstract-Social networking websites and Web 2.0 sites establish interaction between millions of different kinds of people through chat, blogs, newsgroups, news feeds and discussion to express their opinions and sharing of useful knowledge. With these websites subjective data and user-generated content is rapidly increased. The Social networking data is unstructured and noisy in nature, because it is a natural language form. So analyzing and extracting Knowledge from such data are very difficult. Sentiment analysis or opinion mining is the use of natural language processing, text analysis and computational linguistics to identify and extract subjective information in source materials. In this paper, we explain some text and sentiment analysis and its application techniques andwe focus on five machine learning techniques to discover various textual patterns from the social Web.In this paper firstly we proposed a novel opinion mining model for selecting the classifier which gives the highest accuracy, and takes the lowest required time to build the model for movie reviews sentiment analysis. The experimental study showed that Support Vector Machine and Naïve Bayes Multinomial produced the highest accuracy, but the second reduced the time required to build the model much more than the first. Secondly we comparedthe evaluation between the proposed model and the existing model. Keywords-Opinion mining, data pre-processing; feature selection; Machine Learning, Sentiment classification, and Opinion Summarization. I. INTRODUCTION Social networking websites helps people to share useful information with each other. Users who use this websites can establish social relationship with others. This social relationship would make people express their opinions, emotions, feelings, attitudes and knowledge. Sentiment analysis plays very important role for analyzing user contents, their sentiments and emotions, for extracting valuable knowledge automatically. Companies exploit that feature existing in social media web sites to evaluate their own performances and predict new products sales. Sentiment analysis also known Opinion mining. Opinion mining is the process of extracting positive, negative or neutral knowledge from people opinions dataset automatically. Many companies use Opinion mining to analyze product feedback to measure their performance, to predict new product sales. And to help customer to make Purchasing decision. Analyzing user generated content Throw text mining and sentiment analysis by applying both of machine learning techniques which used for text classification using different classification algorithms such as Naive Byes, Decision etc. And Sentiment Analysis by applying other methods such as dictionaries, word lexicons, and word senses etc. The main purpose of this work is firstly to propose an effective framework to perform opinion mining process, and select the best classifier which gives high classification accuracy based on feature selection (FS) technique, the way by which we train and test the model and the least time taken to build the model. In this framework we applied five of the machine learning approaches, Naïve Bayes (NB), Naïve Bayes Multinomial (NBM), support Vector Machine (SVM) with poly kernel, SVM with RBF kernel and RandomForest (RF) for sentiment analysis.Secondly a comparative study between the proposed model and the existing model was performedwe selected Bo Pang, Lillian Lee, and Shivakumar Vaithyanathan 2002 paper as a baseline paper. We will indicate to it in this paper by the baseline paper. In this paper Section II gives background and related work. Section III presents the proposed Framework and Methodology. Machine Learning measurements evaluation will be presented in Section IV. Section V shows Results and discussion. Section VI illustrates Conclusion and future work. II.BACKGROUND AND RELATED WORK In recent years, the research area of sentiment analysis and opinion mining has emerged this section summarize some researcher efforts in this area. A. Literature review In [5] a unified framework which integrates Information retrieval (IR), Opinion Mining (OM) and Text Summarization (TS) together is proposed. In [15] machine learning techniques effectiveness of the sentiment classification problem on movie