ABSTRAKSI: Recommender system (RS) secara umum memiliki banyak manfaat yang penting bagi user dalam interaksinya sehari-hari dengan aplikasi berbasiskan web, khususnya bidang e-commerce. User dihadapkan kepada persoalan mengenai pemilihan produk, content, film, buku dan lain-lain untuk memenuhi kebutuhannya. Oleh karena itu, kemampuan mesin perekomendasi tersebut dibuat dengan tujuan membantu user dalam mengolah dan menghasilkan output system yang dapat menyesuaikan kebutuhan mereka. Recommender System based on Content Filtering dengan algoritma Apriori merupakan salah satu pendekatan RS yang diimplementasikan dalam Tugas Akhir ini. Apriori tergolong algoritma berbasiskan asosiasi rule yang mudah diimplementasikan karena sederhana. Apriori bertugas menambang kombinasi feature dimana feature ini merupakan representasi gambaran item film dataset MovieLens yang dipakai dalam tugas akhir ini yaitu berupa genre film. Rule yang dihasilkan ditujukan untuk proses penambangan genre setiap item film yang disukai user. Hasilnya berupa asosiasi paket genre yang harus terbungkus dalam sebuah film dan hal ini menjadi tolak ukur proses rekomendasi terhadap item film dalam data testing.
Hasil pengujian menunjukkan kinerja performansi sistem yang baik, dibuktikan dengan precision, recall dan f-measure yang meningkat (optimum) seiring sejalan saat menggunakan minimum support = 2%, minimum confidence = 70%, dan data training = 70%. Pada posisi ini juga, rule yang dihasilkan paling sedikit sehingga kinerja mesin tidak hanya baik, akan tetapi berlangsung cepat.
Kata Kunci : Recommender system based on Content Filtering, Apriori, association rule,ABSTRACT: Recommender system (RS) in general has many benefits that are important to the user in everyday interactions with web-based applications, especially in the field of e-commerce. Users faced with problems related to product selection, content, movies, books and others to meet their needs. Therefore, the ability of recommender machine was made with the purpose of assisting users in the process and generating output system that can adapt to their needs. Recommender System based on Content Filtering with Apriori algorithm is one of the RS approach, implemented in this thesis. Apriori association rule-based algorithm classified an easy to implement because it is simple. Priori assigned to mine the combination of this feature where the feature film is a representation of the item description MovieLens dataset used in this thesis is a film genre. Generated rule mining process is intended to genre films which each item is preferred by users. The result is a genre package associations must be wrapped in a film and this is a barometer of the recommendations of film items in the testing data.
The test results show good performance of the system performance, evidenced by the precision, recall and f-measure increased (optimum) go hand in hand while using minimum support = 2%, minimum confidence = 70%, and the training data = 70%. In this position, the rule which generated at least so that engine performance is not only good, but going fast.
Keyword: Recommender system based on Content Filtering, Apriori, association rule,