Turkish Journal of Physiotherapy and Rehabilitation; 32(3) ISSN 2651-4451 | e-ISSN 2651-446X www.turkjphysiotherrehabil.org 284 INTELLIGENT AND EFFECTIVE CROP YIELD PREDICTION USING MACHINE LEARNING TECHNIQUES FOR CLOUD-BASED BIG DATA ANALYTICS Kanagaraj Narayanasamy 1 , R. Pandi Selvam 2 , P. Pandi Selvi 3 , K. Selvan 1 , M. Ilayaraja 4,5 1 Assistant Professor, Department of Computer Applications, J.J College of Arts and Science (Autonomous), Affiliated to Bharathidasan University, Tiruchirappalli, Sivapuram, Pudukkottai, India. K.N: (kanagaraj.n.in@ieee.org),K. S (kselvanjj75@gmail.com) 2 Assistant Professor & Head, PG Department of Computer Science, Ananda College (Affiliated to Alagappa University, Karaikudi), Devakottai, India. (pandiselvamraman@gmail.com) 3 Department of Computer Science, Mangayarkarasi College of Arts and Science for Women, Madurai, Tamil Nadu 625018, India. (selvim11215@gmail.com) 4,5 School of Computing, Kalasalingam Academy of Research and Education, Krishnankoil, India. (ilayaraja.m@klu.ac.in) 5 Corresponding author: M. Ilayaraja, (ilayaraja.m@klu.ac.in) ABSTRACT Agriculture has been an obvious target for big data in environmental conditions, soil variability, input conditions, additives, and commodity prices make it ideal for farmers to use the information and assist in decision making. The reduction of the data is actually obtained using the analysis of the kernel principal component analysis (KPCA). In addition, reducing the graph involves two main procedures, such as Mapper and reducer. When the type of soil is determined on the side of the Mapper, the investigation process takes place on the side of the reducer. Necessary agricultural decisions soil type is classified using the Kernel Fuzzy C Means Clustering (KFCM) and it offers high accuracy gathering. In addition, the innovative technique takes into account the recommendations and a prediction of crop yields, improving the network of proposed Artificial Neural Network with Elephant Herding Optimization (EHO-ANN). In the study, analyzing large-scale crop, soil, and climate data and new non-experimental data will improve production and make agriculture more resistant to climate change. Keywords: Artificial Neural Network, Kernel Principal Component Analysis, Elephant Herding Optimization, Crop Yield Prediction. I. INTRODUCTION Agricultural growers around the world emphasize the need for an exciting means of predicting and improving crop growth. The requirement for coordinated harvest control with a precise boat shaping technique is particularly felt in the provincial network [1]. Agricultural is one of the most important economic sectors in India. It does an essential job of improving the country and supporting it. The degree of agricultural activity can be reduced due to ingredients, for example, unusual rainfall, environmental changes, excessive use of pesticides, etc. [2]. Forecasting crop yields ahead of time can support livestock farmers and government agencies to store, work, help with less value, import/market, etc. [4]. The main objective of this research is to provide a philosophy for the cultivation of gardens that depend on the time atmosphere and the creation of information [2]. Determining yields is an essential agricultural matter. The most experienced breeders have used it to predict their presentation from past performance meetings. In this way, for such distributions of information in the definition of crops, there are various processes or calculations, and using these calculations, we can predict crop yields [3]. Farmers face the need to settle on extreme choices with the most appropriate method to stay profitable and rational with environmental change and the monetary burden of the market. Providing accurate and convenient data; for example, climate, soil, manure use, pesticide use can help breeders find the best choice for their yield. This will help them achieve higher harvest efficiency if conditions are optimal or help them with reduced misery due to competitive harvesting conditions [5].