Development of a Content-Based Filtering Recommendation System for Indonesian Batik Using Motif Features and Regional Origins
Abstract
The diversity of batik motifs and fragmented product information can make it difficult for users to identify motifs that match their preferences. This study develops BatikMatch, a content-based batik motif recommendation system using term frequency-inverse document frequency (TF-IDF) weighting and cosine similarity. The study followed a research-and-development approach with a prototyping model. The dataset consisted of 75 motifs collected from four micro, small, and medium-sized enterprises and batik shops in East Java. Seven attributes (motif name, category, regional origin, color, production technique, philosophy, and intended style of use) were combined into textual descriptions and processed through case folding, tokenization, stopword removal, and stemming. TF-IDF weighting produced a 75 x 172 feature matrix, and cosine similarity yielded a 75 x 75 similarity matrix. With Sparkling as the reference motif, the four top-ranked recommendations were Remo Surabayan (54.30%), Skena Surabaya (38.15%), Tugu Alun-Alun Mojokerto (36.05%), and Belimbing Lining Limo (35.26%). Black-box testing of the user and administrator interfaces showed that the tested functions behaved as reported. These findings demonstrate that BatikMatch can rank batik motifs by similarity between textual attributes. However, the relevance of its recommendations has not yet been assessed through independent user judgments or ranking metrics.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Amallia Berliany Putri, Dodik Arwin Dermawan, I Gde Agung Sri Sidhimantra, Rosita

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Abstract views: 34
,
PDF Downloads: 7


