Open AccessDesigning of Recommendation Engine for Recyclable Waste Mobile App
Author(s) -
Rio Yunanto
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/662/2/022100
Subject(s) - collaborative filtering , computer science , similarity (geometry) , upload , recommender system , data mining , information retrieval , service (business) , mobile apps , global positioning system , preference , artificial intelligence , world wide web , mathematics , statistics , telecommunications , economy , economics , image (mathematics)
The objective of this research is to design a recommendation engine for Pilah Matur App. The recommendation engine in this research used a combination of CF (Collaborative Filtering) and LBS (Location Based Service). Collaborative filtering performs data filtering based on the similarity of user characteristics so that it is able to provide new information to other users because CF provides information based on a pattern of one user group that has a similarity. The preference for recycled material used the nearest neighbor similarity method based on GPS coordinates where the recycled material was uploaded by the recyclable waste donor or volunteer then compared to the location of the recyclable waste taker. The recommendations proposed by collaborative filtering methods could be measured for accuracy using Mean Absolute Error (MAE). The results of the MAE calculation in the user-based CF method, the App Maturity dataset has an MAE value of 0.33. While the item-based CF method gets an MAE value of 0.17 using the same dataset. The results of testing the CF method show that in user-based collaborative filtering the prediction errors are more than in item-based collaborative filtering. The results of the CF method recommendations then sorted based on the closest distance by the LBS method.