Determining The Effect of Some Mechanical Properties on Color Maturity of Tomato With K-Star, Random Forest and Decision Tree (C4.5) Classification Algorithms

This study was conducted in order to determine the effect of the mechanical properties such as maximum force at the skin rupture point, energy at the skin rupture point and the skin firmness on color maturity of tomato by supervised learning algorithms of data mining. In the present study, a total of 88 tomato samples were used, and color measurements for each tomato in 4 different equatorial regions were performed and a total of 352 color measurement units were used. In the classification processes performed according to these mechanical properties, K-Star, Random Forest and Decision Tree (C4.5) algorithms of data mining were utilized, and in the comparison of comprising classification models, Root Mean Square Error (RMSE), Mean absolute error (MAE), Root relative squared error (RRSE) and Relative absolute error (RAE) values, which are some of the criteria of error variance, were considered to be low, while the classification accuracy rate was considered to be high. As a result of the comparison made, the classification model formed according to K-Star instance-based algorithm [MAE: 0.004, RMSE: 0.006, %RAE: 1.73, %RRSE: 1.70] has been found to be a better classifier compared to the others. With the classification made according to K-Star algorithm, the maximum force at the skin rupture point on the degree of maturity of tomato and the skin firmness were found to be green, light red, and their effects are non-significant during the color conversion periods, and found significant during other periods while the energy at the skin rupture point is only pink and has been to be significant during the color conversion stages and non-significant during other stages.