Modeling the data collected from smart greenhouses with data mining methods

The increasing population of the world causes the need for food to increase. In order to meet the needs of the people and to offer better quality products, those working in the agricultural sector keep up with the innovations. Developing technology has brought smart agriculture applications with it. In this way, farmers were provided with the opportunity to continue production in smart greenhouses. The aim of the study is to examine the effect of the relationship between irrigation water and average internal temperature in tomato production on the tonnage taken in a smart greenhouse using data mining methods. In addition, algorithms that provide the closest results to reality were determined from the models created within the data of the last four years. In line with the stated purpose, modeling was done using data mining methods, random forest, k-nearest neighbors, artificial neural networks and decision trees algorithms, using the production data obtained from the smart greenhouse. The production data used in the study was obtained from a business using the smart greenhouse application, which started production in Afyon's Sandıklı district in 2004. The data in question is tomato production data between March 2019 and December 2022. Correlation analysis was performed using these data, and irrigation and average internal temperature values were examined in the distribution graph. As a result of the study, the positive effect of irrigation water and average internal temperature factors, which directly affect the yield of tomatoes grown in the smart greenhouse, on production efficiency was confirmed. In addition, as a result of the modeling done separately for each year, different algorithms gave the closest result to reality with the least error. Enterprises engaged in production in the field of agriculture can increase product efficiency by analyzing production data with data mining methods.