Scholarlink

<p>Estimating the quantity of calories in a particular food item has become a common and required practice for the majority of individuals who are attempting to maintain a balanced diet. Accurately predicting the quantity of calories not only provides valuable information regarding one’s diet but also shows whether such a diet is suitable or should be altered. In this paper, research is presented using a Machine Learning (ML) algorithm known as the Multi-Layer Perceptron Regression (MLPR) that is enhanced through the synergy of Gorilla Troop Optimization (GTO) and Prairie Dog Optimization (PDO). Our primary aim was to enhance the accuracy of the calorie predictions as well as reduce error values. To do this, we combined the MLPR model with the said optimization techniques to devise new hybrid schemes that were improved in their performance. Upon exhaustive comparison of output from these hybrid schemes at varying levels of analysis, the research indicated that, at the level of the first layer, the MLGT model recorded a figure of 0.933 in the coefficient of determination (R²), an indication of its relatively poor performance. Contrastingly, the MLPD model yielded a greater R² value of 0.945, reflecting its enhanced efficiency in calorie content prediction. These outcomes highlight the strengths of applying optimized hybrid schemes to dietary prediction and analysis.</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

Hybrid optimization-based predictive techniques for estimating food caloric Content

  • 期刊论文

  • 发表日期：

    2025年10月21日

查看全文
  • Gui Wu,
  • Zejin Liu,
  • Ye Liu

摘要

Estimating the quantity of calories in a particular food item has become a common and required practice for the majority of individuals who are attempting to maintain a balanced diet. Accurately predicting the quantity of calories not only provides valuable information regarding one’s diet but also shows whether such a diet is suitable or should be altered. In this paper, research is presented using a Machine Learning (ML) algorithm known as the Multi-Layer Perceptron Regression (MLPR) that is enhanced through the synergy of Gorilla Troop Optimization (GTO) and Prairie Dog Optimization (PDO). Our primary aim was to enhance the accuracy of the calorie predictions as well as reduce error values. To do this, we combined the MLPR model with the said optimization techniques to devise new hybrid schemes that were improved in their performance. Upon exhaustive comparison of output from these hybrid schemes at varying levels of analysis, the research indicated that, at the level of the first layer, the MLGT model recorded a figure of 0.933 in the coefficient of determination (R²), an indication of its relatively poor performance. Contrastingly, the MLPD model yielded a greater R² value of 0.945, reflecting its enhanced efficiency in calorie content prediction. These outcomes highlight the strengths of applying optimized hybrid schemes to dietary prediction and analysis.