Demand Prediction of Consumer Intention to Buy Edible Items Using Machine Learning Techniques

Agriculture is a process of simplifying the food chain in nature and redistributing the resources for animals and plant production. It is the process of nurturing livestock and growing crops. It consists of preparing animal and plant items for human consumption and distributing them to marketplaces. A maximum number of textiles and food items in the world are produced by agriculture. In India, the primary source of income is agriculture. Various companies buy crops from the farmers and directly sell them to the consumers. As per a study by “The India Express”, there are 90–150 million farmers who rely on agriculture to run the expenses of their daily lives. Each day there is a huge amount of vegetable waste as their demand doesn't get predicted due to which a volume of vegetables gets rotten. So to deal with this issue, there is a need for the development of an automated model which can accurately predict the demand for vegetables for the next day based on the previous data. In this work, a framework has been developed for predicting the demand for okra and tomato for the next day, so that the farmers won't pluck too many vegetables and also keep track of daily sales. We have used four machine learning models namely, linear regression, decision tree, random forest, and logistic regression for identifying the demand for vegetables. In this experiment, two parameters %error and accuracy have been taken which shows that the decision tree model performs best with the highest accuracy of 99.62% and logistic regression has shown the least accuracy at 85.62%. This work can help various edible-based eCommerce companies for predicting the demand for the items for the next day based on the previous data.