<![CDATA[The growth in the number of cinema audiences is increasing in line with the large number of films being produced. Various films with plot stories, genres, and film themes that are similar or different have enlivened the industrial market from overseas to domestic film. Of the many films produced, it makes potential viewers confused and difficult to find and determine what film to watch next so that they spend more time looking for films. Some people use the features provided on some sites to search for movies to decide which movie to watch. Everyone has different tastes and tends to choose to watch movies that are similar to the movies he likes. One way to get the right information about a film is a recommendation system. Each film has some information in the form of different genre films and synopsis films. In this study, to obtain the recommendation results using a content based filtering algorithm by looking for the similarity in weight of the terms in the bag of words result of pre-processing film synopsis and film title. The weighting is carried out using the TF-IDF method which has been normalized. Then the weighting results will go through the cosine similarity stage to look for similarities based on weights and end with filtering based on genre. Based on the results of tests carried out by involving three participants with a total number of films as many as 4000 film titles, the accuracy value is obtained using the mean average precision @K (MAP @ K) is 0.823254 for the single query type and 0.7500556 for the multiple seed query type. From these results, it is found that the single query type produces better recommendations than the multiple seed query type.]]>
Copyrights © 2021
