Application of Artificial Neural Networks to predict the final fruit weight and random forest to select important variables in native population of melon (Cucumis melo L.)

- The results obtained highlight the obvious usefulness of Artificial Neural Networks, which could now be used in a prospective evaluation for application of the technique to predict fruit weight.
- Artificial Neural Network model successfully predicted fruit weight.
- Applications of random forest method for variable selection showed importance of the flesh diameter in melon.

The estimation of the relation between the inconstant factors can be highly helpful for calculating the amount of variation of a particular character with respect to others. This paper aims to study the effects of different agronomic and phenologic factors on the total mass of melon fruit produced. The agronomic and phenologic factors which were considered during the study included, plant length, fruit weight, fruit length, fruit width, number of fruits per each plant, number of days to flowering, number of days to maturity, number of days to fruit formation, fruit cavity diameter and flesh diameter were the other characters under study. During the study, every plant was taken as a self-sustaining unit. The study explains a procedure to foretell the yield of melon by applying the Artificial Neural Networks or ANNs as a displaying instrument. In the study the accession Firoozi was calculated with high accuracy and efficacy (R2 = 87%, EMP = 2.21 and MSD = 1.66). RF yield variable importance measures for each candidate predictor and in this study flesh diameter examined as effective variable in identifying the true predictor among the candidate predictors.