EUVIMEDEUVIMED entdecken
Crossref

Enhanced Fashion Demand Forecasting Using a Hybrid CNN LSTM Deep Learning Framework

Hemangini V. Patel, Kirit J. Modi, Khyati A. Prajapati,, Ekta V. Patel

SPU - Journal of Science, Technology and Management Research · 01.07.2026 · 121-131

DOI: 10.63766/spujstmr.26.000079 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SPU - Journal of Science, Technology and Management Research, online erschienen am 01.07.2026. Verfasst von Hemangini V. Patel und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate demand forecasting in the fashion retail industry is a critical yet challenging task due to rapidly changing consumer preferences, short product life cycles, seasonal variability, and the strong influence of visual product attributes on purchasing decisions. Traditional statistical forecasting methods such as ARIMA and exponential smoothing are limited by linear assumptions and therefore fail to effectively model the complex, non-linear patterns present in fashion sales data. Although recent advances in deep learning have shown promising results, single-model approaches often struggle to simultaneously capture spatial and temporal characteristics of fashion demand. To address these limitations, this paper proposes a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) deep learning framework for enhanced fashion demand forecasting. The CNN component is employed to extract high-level spatial and visual features from fashion product images and associated attributes, while the LSTM component models long-term temporal dependencies from historical sales and inventory time-series data. Experiments are conducted using the Fashion Product Images Dataset and the Retail Store Inventory Forecasting Dataset obtained from Kaggle. The performance of the proposed model is evaluated using standard forecasting metrics, including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Experimental results demonstrate that the proposed CNN–LSTM framework achieves superior forecasting accuracy compared to traditional statistical methods and standalone deep learning models. The proposed approach offers a scalable and effective solution for inventory planning and demand management in the fashion retail domain.

Zitieren

APA

Patel, H. V., Modi, K. J., Prajapati,, K. A., & Patel, E. V. (2026). Enhanced Fashion Demand Forecasting Using a Hybrid CNN LSTM Deep Learning Framework. SPU - Journal of Science, Technology and Management Research, 121-131. https://doi.org/10.63766/spujstmr.26.000079

Vancouver

Patel HV, Modi KJ, Prajapati, KA, Patel EV. Enhanced Fashion Demand Forecasting Using a Hybrid CNN LSTM Deep Learning Framework. SPU - Journal of Science, Technology and Management Research. 2026:121-131. doi:10.63766/spujstmr.26.000079

BibTeX
@article{Patel2026,
  author = {Patel, Hemangini V. and Modi, Kirit J. and Prajapati,, Khyati A. and Patel, Ekta V.},
  title = {Enhanced Fashion Demand Forecasting Using a Hybrid CNN LSTM Deep Learning Framework},
  journal = {SPU - Journal of Science, Technology and Management Research},
  year = {2026},
  pages = {121--131},
  doi = {10.63766/spujstmr.26.000079}
}
Kostenlos starten →