Application of Machine Learning to Study the Agricultural Mechanization of Wheat Farms in Egypt

Abstract

Agricultural production can achieve sustainability by appropriately applying agricultural mechanization, especially in developing countries where smallholding farmers lack sufficient agricultural machinery for their farming operations. This paper aimed to study the extent to which small-, medium-, and large-scale farms in the Delta of Egypt use agricultural mechanization in their wheat crop farming operations. K-means clustering was used to aggregate and analyze the scenarios implemented by farmers for wheat cultivation so as to suggest guidelines for each cluster of farmers on how to mechanize their indoor wheat agricultural operations to maximize production. The study is divided into two parts: Firstly, data were collected regarding the percentage of small, medium, and large farms; the cultivated area of wheat crops in small-, medium-, and large-scale farms; and the size of tractors, as an indicator of the mechanization available in the governorates of Egypt’s Delta. Secondly, data were collected through a questionnaire survey of 2652 smallholding farmers, 328 medium-holding farmers, and 354 large-holding farmers from Egypt’s Delta governorates. Based on the surveyed data, 14, 14, and 12 scenarios (indexes) were established for small-, medium-, and large-scale farms, respectively, related to various agricultural operations involved in wheat crop production. These scenarios were analyzed based on the centroids using K-means clustering. The identified scenarios were divided into three clusters for the three levels of farms. The data obtained showed the need for smallholding farmers to implement mechanization, which could be achieved through renting services. These findings, if implemented, would have huge social and economic effects on farmers’ lives, in addition to increasing production, saving time and effort, and reducing dependence on labor.