Textile weaving dataset for machine learning to predict rejection and production of a weaving factory

Toufique Ahmed; Shihab Uddin

doi:10.1016/j.dib.2023.108995

Textile weaving dataset for machine learning to predict rejection and production of a weaving factory

Data Brief. 2023 Feb 21:47:108995. doi: 10.1016/j.dib.2023.108995. eCollection 2023 Apr.

Authors

Toufique Ahmed¹, Shihab Uddin²

Affiliations

¹ Department of Textile Engineering, National Institute of Textile Engineering and Research (NITER), Dhaka, Bangladesh.
² Senior Manager (weaving), Evince Textiles Ltd., Gazipur, Bangladesh.

Abstract

Weaving is one of the most popular fabric manufacturing techniques. The weaving process consists of 3 major stages: warping, sizing, and weaving. The weaving factory henceforth involves a lot of data. But unfortunately, there is no attempt to utilize machine learning or data science in weaving production. Although a variety of scopes are there to implement statistical analysis, data science, and machine learning. The dataset was prepared by using the daily production report for 9 months. The final dataset contains 121,148 data with 18 parameters. Whereas the raw data contains the same number of entries with 22 columns. The raw data needs substantial work to combine the daily production report, treat the missing values, rename columns, and feature engineering to derive EPI, PPI, warp, weft count values, etc. The complete dataset is stored at https://data.mendeley.com/datasets/nxb4shgs9h/1. It is further processed to get the rejection dataset which is stored at https://data.mendeley.com/datasets/6mwgj7tms3/2. The future implementation of the dataset is to predict the weaving waste, investigate the statistical relations among various parameters, production prediction, etc.

Keywords: Machine learning; Weaving production data; Woven fabric rejection; Woven fabrics.