Rose Multiclassification: Harnessing Hybrid CNN and Random Forest Model

The study of roses classification has the goal to collect the data, to classify, to preprocess the data and finally to extract it's features according to rose's uniqness or dissimilarities using color and shape features and their combinations. This research falls under the hybrid approach on the regions of Punjab and some parts of Haryana. The accuracy fully achieved is 6.85. The maximum accuracy in Class 1,2,3,5,6,7 is 0.98 and Minimum accuracy is class 4 is 0.97. The macro average for precision, recall and F1-Score is 91.76%, 91.75% and 91.75% respectively. Weighted Average for precision, recall and F1-Score is 91.84%, 91.82% and 91.82% respectively. Micro Average for precision, recall and F1-Score is 91.82%, 91.82% and 91.82% respectively. This research includes 7 classes. This classification helps to provide support and care to many different garden roses or hybrid roses and to have a keen eye on their growth habits. After comparing the results of the criterion their varieties, types and its features for multi classification of roses, the conclusion is to classify more about roses and to provide assistance and care.