PREDICTING DEMAND FOR COTTON YARNS

Abstract

Predicting demand for fashion products is crucial for textile manufacturers. In an attempt to both avoid out-of-stocks and minimize holding costs, different forecasting techniques are used by production managers. Both linear and non-linear time-series analysis techniques are suitable options for forecasting purposes. However, demand for fashion products presents a number of particular characteristics such as short life-cycles, short selling seasons, high impulse purchasing, high volatility, low predictability, tremendous product variety and a high number of stock-keeping-units. In this paper, we focus on predicting demand for cotton yarns using a non-linear forecasting technique that has been fruitfully used in many areas, namely, random forests. To this end, we first identify a number of explanatory variables to be used as a key input to forecasting using random forests. We consider explanatory variables usually labeled either as causal variables, when some correlation is expected between them and the forecasted variable, or as time-series features, when extracted from time-related attributes such as seasonality. Next, we evaluate the predictive power of each variable by means of out-of-sample accuracy measurement. We experiment on a real data set from a textile company in Spain. The numerical results show that simple time-series features present more predictive ability than other more sophisticated explanatory variables.

Authors and Affiliations

Francisco SALAS-MOLINA, Pablo DÍAZ-GARCÍA

Keywords

Related Articles

THERMOGRAPHIC ANALYSIS OF THE RICOMA 2 HEAD EMBROIDERY MACHINE

The purpose of this paper is to quickly and accurately identify the points representing potential defects in 2 head embroidery machines by using infrared thermography. Thermography measurements were performed with the FL...

PHOTOCATALYTIC EFFICIENCY OF N-TIO2 APPLIED ON COTTON KNIT – PART 1

The main aim of the research work is the development of photocatalytic textiles materials by treating them with TiO2 doped with nitrogen. Also, the research was focused on the nitrogen-doped titanium dioxide (N-TiO2) for...

FE-SEM COMPARATIVE STUDY ON SURFACE MODIFICATION OF WOOL FIBER AFTER DIFFERENT CHEMICAL TREATMENTS

Wool surface comprehends numerous scales which are responsible of certain undesirable behavior of this fiber during its use and maintenance. One of the most significant issues is related to shrinkage, caused during washi...

MAINTENANCE PLANNING OF THE SEWING NEEDLES OF SIMPLE SEWING MACHINES

The effectiveness of simple sewing machines can be increased through the planning of predictive maintenance activities. The monitoring of the technical condition of the sewing needles of simple sewing machines was based...

USEBILITY OF HYDROGELS IN ADSORPTION TECHNOLOGHY FOR REMOVAL OF HEAVY METAL AND DYE

Heavy metals and Dyes are very toxic and nonbiodegradable in waste waters to cause adverse health effects in human body and to induce irreversible pollution. Adsorption offers many potential advantages for removal of tox...

Download PDF file
  • EP ID EP216645
  • DOI -
  • Views 63
  • Downloads 0

How To Cite

Francisco SALAS-MOLINA, Pablo DÍAZ-GARCÍA (2017). PREDICTING DEMAND FOR COTTON YARNS. Annals of the University of Oradea. Fascicle of Textiles, Leatherwork, 0(1), 97-102. https://europub.co.uk/articles/-A-216645