Scholarlink

<p>Prediction of agricultural commodity prices is challenging due to the intricate interaction of multiple influencing factors which poses limitations for traditional statistical tools. The machine learning (ML) techniques e.g. artificial neural network (ANN), random forest (RF), support vector regression (SVR) are widely used in the domain of time series forecasting to capture the nonlinearity. On the other hand, Fuzzy technique has proven its potential in field of forecasting for ambiguous and vague data. Traditional stochastic and ML models often fail to capture the noise, uncertainty, incompleteness, non-linearity, and complexity presence in data and thus, cannot give accurate predictions. To improve forecasting accuracy by handling uncertainty and imprecise information in real-world datasets, this study proposes a fuzzy-based machine learning model (FuzzyML). It combines fuzzy logic with machine learning techniques. The proposed models i.e. Fuzzy-ANN (F-ANN), Fuzzy-RF (F-RF) and Fuzzy-SVR (F-SVR) have been developed by incorporating advantages of fuzzy relations in ML models by using membership and non-membership values through intuitionistic fuzzy c-means technique. To examine the predictive performance of FuzzyML model, India’s monthly wholesale prices (Rs/q) of 18 agricultural commodities namely Rice, Wheat, Maize, Gram, Urad, Lentil, Mung, Arhar, Potato, Onion, Tomato, Soyabean, Sunflower, Groundnut, Mustard, Turmeric, Coriander, and Cumin have been used during the period January 2010 to December 2022. The accuracy of prediction of proposed FuzzyML model is evaluated in comparison with other benchmark models i.e. Autoregressive Integrated Moving Average (ARIMA), ANN, SVR, RF and also Fuzzy ARIMA (F-ARIMA) using different accuracy measures. To gain a comprehensive understanding of all accuracy measures, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) has been applied. The study revealed that incorporating fuzzy logic to ML model increases the predictive performance of the models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

FuzzyML: fuzzy based machine learning models for predicting agricultural prices

  • 期刊论文

  • 发表日期：

    2025年6月30日

查看全文
  • Anita Sarkar,
  • Ranjit Kumar Paul,
  • Md Yeasin,
  • Ankit Kumar Singh,
  • A. K. Paul

摘要

Prediction of agricultural commodity prices is challenging due to the intricate interaction of multiple influencing factors which poses limitations for traditional statistical tools. The machine learning (ML) techniques e.g. artificial neural network (ANN), random forest (RF), support vector regression (SVR) are widely used in the domain of time series forecasting to capture the nonlinearity. On the other hand, Fuzzy technique has proven its potential in field of forecasting for ambiguous and vague data. Traditional stochastic and ML models often fail to capture the noise, uncertainty, incompleteness, non-linearity, and complexity presence in data and thus, cannot give accurate predictions. To improve forecasting accuracy by handling uncertainty and imprecise information in real-world datasets, this study proposes a fuzzy-based machine learning model (FuzzyML). It combines fuzzy logic with machine learning techniques. The proposed models i.e. Fuzzy-ANN (F-ANN), Fuzzy-RF (F-RF) and Fuzzy-SVR (F-SVR) have been developed by incorporating advantages of fuzzy relations in ML models by using membership and non-membership values through intuitionistic fuzzy c-means technique. To examine the predictive performance of FuzzyML model, India’s monthly wholesale prices (Rs/q) of 18 agricultural commodities namely Rice, Wheat, Maize, Gram, Urad, Lentil, Mung, Arhar, Potato, Onion, Tomato, Soyabean, Sunflower, Groundnut, Mustard, Turmeric, Coriander, and Cumin have been used during the period January 2010 to December 2022. The accuracy of prediction of proposed FuzzyML model is evaluated in comparison with other benchmark models i.e. Autoregressive Integrated Moving Average (ARIMA), ANN, SVR, RF and also Fuzzy ARIMA (F-ARIMA) using different accuracy measures. To gain a comprehensive understanding of all accuracy measures, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) has been applied. The study revealed that incorporating fuzzy logic to ML model increases the predictive performance of the models.