Design a personalized Recommender system based on sentiment analysis on social media (Case study: banking system)

Customer retention is one of the most important issues of any organization and finding a way to retain and maintain the customer is one of the critical needs of any organization. The main purpose of the present study in the field of machine learning is to focus on the problem of correctly identifying customer needs with a method based on extracting opinion and sentiment analysis and quantifying customers' sentiment orientation. In the other word, the main issue is to design a Recommender System to provide appropriate services in accordance with customer satisfaction, sentiment, and experiences. The proposed method is that customers' opinions and experiences are obtained by evaluating tweets containing hashtags with the titles and headings of banking services as statistical population, and after revision, it results in providing correlation scores in terms of people's sentiment score due to the tweets, cosine similarity and reliability, consideration of relevant characteristic groups as well as recorded ideas in the training and testing process, in the form of submitting personalized offer to receive banking services. In order to represent a recommending solution, suitable classification methods are used along with opinion mining methods and proper validation approach as well, and the terminal designed system with a little error will take steps to provide personalized services as well as help banking system. Since there is no thorough provision of banking services tailored to the customers’ situation, so in this regard, the mentioned system will be extremely beneficial.