Recommender systems are information retrieval tools helping users in their informationseeking tasks and guiding them in a large space of possible options. Many hybridrecommender systems are proposed so far to overcome shortcomings born of purecontent-based (PCB) and pure collaborative ltering (PCF) systems. Most studies onrecommender systems aim to improve the accuracy and eciency of predictions. Inthis thesis, we propose an online hybrid recommender strategy (CBCFdfc) based oncontent boosted collaborative ltering algorithm which aims to improve the predictionaccuracy and eciency. CBCFdfc combines content-based and collaborative characteristicsto solve problems like sparsity, new item and over-specialization. CBCFdfc usesfuzzy clustering to keep a certain level of prediction accuracy while decreasing onlineprediction time. We compare CBCFdfc with PCB and PCF according to predictionaccuracy metrics, and with CBCFonl (online CBCF without clustering) according toonline recommendation time. Test results showed that CBCFdfc performs better thanother approaches in most cases. We, also, evaluate the eect of user-specied parametersto the prediction accuracy and eciency. According to test results, we determineoptimal values for these parameters. In addition to experiments made on simulateddata, we also perform a user study and evaluate opinions of users about recommendedmovies. The results that are obtained in user evaluation are satisfactory. As a result,the proposed system can be regarded as an accurate and ecient hybrid online movierecommender.Keywords: hybrid recommender system, content-based systems, collaborative lteringsystems, fuzzy clustering |