HYBRID CNN-LSTM MODEL: RAINFALL ANALYSIS AND PREDICTION FOR KARNATAKA REGION

Purpose Rainfall prediction is necessary for harvesting crops throughout the year, agriculture yields crops based on the farmer's ability to work in a specific field for particular crop fertilization. This idea was not alone necessary to predict the crop's yield. Seed firms regularly screen how efficiently plant varieties grow in a particular setup. Thirdly to predict agricultural produce is critical to solving emerging concerns for food security in the phase of global climatical changes. Accurate prediction of forecasts assists farmers so they can take more economical and cost-management decisions. They also enhance the prevention of famine. This results in protecting farmers' efficiency and productivity to reduce the risks associated with environmental gain. Design/Methodology/Approach- This paper proposes a Hybrid machine learning model for efficient prediction of rainfall, this makes a solution that is used for encoding them to create solutions abruptly. A major part of this work here is focused on generating a solution for fitness to meet the highest accuracy. This algorithm works efficiently for the given input data. This algorithm tries to meet the necessary requirements until an optimal analysis is carried out to contribute to maximum accuracy for rainfall prediction. Findings- The performance of our proposed Hybrid Algorithm is compared with existing algorithms such as linear regression, logistic regression, and KNN. The proposed Hybrid Algorithm convolutional neural networks with LSTM (Long short-term memory model) with First-order optimization Algorithm which works along the gradient-descent algorithm based on different metrics like Accuracy, Sensitivity, Specificity, F-score, and found maximum performance. Originality/value- This paper proposes a Hybrid algorithm CNN_LSTM along with this a first-order optimization algorithm that functions based on the gradient descent method. The results found using the proposed algorithm are plots. A comparative analysis is carried out using various results obtained to achieve high performance to solve different constraints.