Enhancing Cucumber Leaf Disease Severity Prediction through Convolutional Neural Networks and Random Forest Integration

The abstract functions as a succinct summary of the main conclusions drawn from the study. By utilizing the extensive classification performance metrics shown in the attached table, the abstract highlights the accuracy and efficiency of the model in predicting the various severity levels of Cucumber Mosaic Virus (CMV) infections on cucumber leaves. The model is very good at defining CMV infection levels; the precision, recall, or F1-score metrics show this clearly, with values that are consistently high for all severity levels. Furthermore, the support values provide insightful information about the distribution of the dataset by highlighting the frequency of occurrences for each severity level. The model demonstrated an outstanding 97.59% overall accuracy on a dataset consisting of 4908 photos of cucumber leaves. The Support Proportion boundary further emphasizes how different severity categories are represented within the dataset, which improves the comprehension of the makeup of the data. Additionally, It has been known that the overall accuracy rate is not good, which indicates that many people get it wrong. However, if they look at a weighted average level, they can see that it is quite effective in predicting disease severity caused by CMV. This astute finding not only shows the effectiveness of the recommended methodology but also implies interesting applications in agricultural areas. When diseases become severe, the model has significant potential to provide valuable insights for timely interventions and strategic management methods. Cucumber production will then realize the protection of crop productive yield and improved product quality under the pressure of CMV infection.