期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2016
卷号：4
期号：2
页码：2125
DOI：10.15680/IJIRCCE.2016.0402067
出版社：S&S Publications
摘要：A huge amount of data which is generated by the user is available in the medical forums with the help of online health care social media. This social media helps the patients to play an important role in the medical forums like get the information about medicine, treatment.The patientscan actively interacting with other patients or physicians in the online healthcare discussions.But they can't find the related topics in that medical forum. It is some what not easy to the users. Because the data's in medical forums are in unstructured format. Recommendation systems are basedon content based approach which recommends about discussions to the patients. But the content based approach is not sufficient. Because the patients may have different intentions or they are having interest in different types even if the data matches their interest. By using Na.ve Byes method, there are two tasks are proposed for classifying posts and comments in the medical forums. They are Classification of intentions and Classification of social support types.Inorder to develop the Classifiers, text feature sets and health feature sets are used. To combine Classifiers with different feature sets and also optimizing the classification results, the genetic algorithm is used. For Post Classification, combining text and health feature sets can achieve the highest precision, recall and F1 measure. For Comment Classification, combining the text feature sets can achieve the best result
关键词：Classification and Optimization Techniques;Genetic Algorithm; Naive BayesAlgorithm;Recommender Systems
