Memprediksi Tren Fashion Busana Muslimah

NIA SASTRA PERMATA, Ir. Anna Maria Sri Asih, S.T., M.M., M.Sc., Ph.D., IPM., ASEAN.Eng

2021 | Tesis | MAGISTER TEKNIK INDUSTRI

Fashion muslimah merupakan salah satu trend fashion populer di Indonesia karena mayoritas penduduk muslim. Salalu satu yang termasuk dalam fashion muslimah adalah baju atasan. Trend fashion muslimah selalu berubah - ubah seiring berjalannya waktu yang disebabkan oleh keinginan konsumen. Perilaku konsumen yang cepat bosan dengan fashion populer dan mencari fashion terbaru lainnya, berdampak pada umur fashion menjadi relatif pendek. Hal ini menjadi tantangan dalam industri fashion untuk lebih fleksibel dan responsif kearah perubahan trend fashion terutama fashion muslimah. Salah satunya dengan melakukan prediksi desain fashion muslimah. Untuk mengetahui informasi desain fashion, diperlukan informasi yang spesifik dari konsumen. Text Mining digunakan untuk menggali data informasi berupa kata – kata, lalu mengidentifikasi kata – kata yang merujuk pada desain baju atasan fashion muslimah dan mengklasifikasikan kata – kata yang bermakna sama menjadi fitur yang sesuai dengan desain baju atasan fashion muslimah. Prediksi dilakukan dengan tiga model peramalan yaitu moving average, centered moving average, dan exponential smoothing, nilai error terkecil terpilih menjadi model peramalan yang terbaik. Terdapat enam parameter desain fashion muslimah yang menjadi dasar prediksi yaitu warna, shape of sleeves, style baju, shape of blouse, shape of tunic, dan additional describing attribute. Model prediksi terbaik yaitu centered moving average (CMA). Akurasi validasi model prediksi secara keseluruhan sebesar 83,3% menunjukkan model prediksi cukup baik. Hasil prediksi untuk desain trend fashion muslimah selanjutnya adalah fashion muslimah dengan style blouse; shape of blouse: peplum blouse, simple blouse, dan asimetris blouse; warna: ungu, hijau, oranye, biru muda, teracotta, dan kuning. shape of slevee: bishop, marie, batwing, peasant; additional describing attribute: polos dan motif penuh.

Muslim fashion is one of the popular fashion trends in Indonesia because of the Muslim population. What is included in Muslim fashion is a top shirt. Muslim fashion trends always change over time due to consumer desires. The behavior of consumers who are quickly bored with currently popular fashion and looking for other latest fashions, has an impact on the relatively short age of fashion. This is a challenge in the fashion industry to be more flexible and responsive to changing fashion trends, especially Muslim fashion. In response to the relatively short age of fashion, one of them is by predicting Muslim fashion designs. To see fashion design information, it requires specific information from consumers. Text Mining is used to search for data information in the form of words, then identify words that refer to the design of Muslim women's fashion tops and classify words that are close to the same meaning into features that are in accordance with the design of Muslim fashion tops. Prediction is done by using three forecasting models, namely moving average, centered moving average, and exponential smoothing, the error value is transformed into the best forecasting model. There are six Muslim fashion design parameters that form the basis of the prediction, namely color, sleeve shape, shirt model, shirt shape, tunic shape, and additional describing attributes. The best prediction model is the centered moving average (CMA). The total accuracy of the prediction model validation is 83.3%, indicating that the model prediction is quite good. The prediction results for the next Muslim fashion trend design are Muslim fashion with a blouse style; blouse forms: peplum blouse, simple blouse, and asymmetrical blouse; colors: purple, green, orange, light blue, teracotta, and yellow; slevee forms: bishop, marie, batwing, peasant; Additional description attributes: plain and full motif

Kata Kunci : Fashion, Muslimah, Prediction, Centered Moving Average, Text mining