Prediction and Quantitative Analysis of Yarn Properties from Fibre Properties Using Robust Regression and Extra Sum Squares
Research and development
Authors:
- Fattahi S.
Department of Textile Engineering, University of Yazd, Yazd, Iran - Hoseini Ravandi S. A.
Department of Textile Engineering, Isfahan University of Technology, Isfahan, Iran
 Full text | Abstract: The main aim of this study is the prediction and quantity evaluation of important yarn properties (tensile, unevenness, hairiness and imperfections of yarn) from fibre properties by the robust regression and extra sum squares methods. We used cotton fibre and yarn properties measured by means of an HVI system and Uster tester. Properties of 87 Controlled samples of ring-spun cotton yarn with linear densities ranging from 19.2 to 37.4 tex with twist multiple: atex = 3927.8 (from from 19.2 to 37.4 tex) were used. In this way we selected the effective variables by considering all possible regressions and through the criteria of the mean square error (MSE) and adjusted R2. Optimum equations with appropriate variables and relative importance of various variables were also investigated. After the fit, desirable MSE statistics and large adjusted R2 values were observed. |
Tags:
robust regression, extra sum squares, yarn imperfections, cotton spinning, yarn
quality properties.
Citation:
Fattahi S, Hoseini Ravandi SA. Prediction and Quantitative Analysis of Yarn Properties from Fibre Properties Using Robust Regression and Extra Sum Squares. FIBRES & TEXTILES in Eastern Europe 2013; 21, 4(100): 48-54.