Agriculture is one of the major sources of employment as well as contributor in the GDP of India. Machine learning is the latest technology which can be used to help the agriculture sector. This paper will focus in using the machine learning technique to predicting the wheat crop yield. The regression algorithms which are used in it are simple regression, gradient booster, polynomial regression and random forest. The results of every algorithm are compared with actual results in the last.
weather, soil, use of fertilizers, and seed variety. This indicates that crop yield prediction is not a trivial task; instead, it consists of several complicated steps. Nowadays, crop yield prediction models can estimate the yield, but a better performance in yield prediction is still desirable (Klompenburg et al. 2020).
A. Agricultural Information Extraction 1) Raorane and Kulkarni (2015), used datamining tools in crop management system. They used regression algorithms. The disadvantage is the model is not specified. 2) Kushwaha and Bhattachrya (2015), concluded the method which is helpful in finding the suitable crop according to the land. Agro algorithm is used in this paper. 3) Santra et al. (2016), used artificial neural network, decision tree algorithm and regression analysis to providing the information of crops and help in increasing the yield rate. The negative is method is not clearly specified.
1) Kumar et al. (2015), suggested the method which is helpful in improving the yield of crops.
Classifications are used and the parameters are compared. The demerit is the accuracy and performance is not proper. 2) Babu and Babu (2016),gave method which provide solutions to some farming problems such as water and fertilizers. They have also used the agro algorithm and the accuracy is also the problem in it. 3) Jain et al. (2017), in their paper found the better sequence according to which the crops should be sown so that the maximum yield is extracted. Not only sequence they also used machine learning for irrigation and crop diseases. 4) Djodiltachoumy (2017), used K means algorithms (Clustering) on previous years data and predict yield according to that database. The demerit is they used fewer amounts of data and it is suitable only for association rule. 5) Nigam et al. (2019), have concluded the random forest regression gives the highest yield prediction accuracy. Simple recurrent neural network performs better on rainfall prediction while LSTM is good for temperature prediction.
C. Machine Learning Algorithms 1) Khairunniza-Bejo et al. (2014), defined a method using Artificial Neural Network to help the farmers solving some of their problems. The disadvantage is the proposed method is very time consuming. 2) Ramesh and Vardhan (2015),used multiple linear regression method to analyze and verify the database. The demerit is this method is of less accuracy. 3) Savla et al. (2015), suggested the framework using Normalization, Clustering and
Classification to understand the crop yield rate zones based on attributes. 4) Sindhura et al. (2016), also used multiple linear regression methods to predict and support the decision making in many sectors.
The comprehensive study of literature review revealed that the crop yield estimation and agricultural information extraction from the ancillary data as well as historical data is an open problem. Various machine learning models and other algorithms have been used in past for the yield estimation.
Collected data is preprocessed. There were some ‘NA’ values which are filled by taking average value of the above and below column.
Feature selection is applied to extract important parameters for modeling framework. A process to find correlation between all the parameters is applied and the parameters which were not affecting the crop yield are eliminated. Image of correlation is given below:
1. Results of applied machine learning algorithms are compared to evaluate the model. The table of results are given below in table 2: 2. The representation of all the predicted values and actual values from year 2011 to 2018 is also given below in line and bar graph:
3. The table of performance evolution measures such as Mean Absolute Error, Mean Squared Error, Root Mean Squared Error and Mean Absolute Percentage Error of applied algorithm is given below in table 3:
709.744 597,836.8 13 773.199 0.144 570.942 440,162.5 65 663.447 0.115
Regression 583.375 452,351
.375 672.571 0.118 986.935 1,102,9 40.261 1050.21 0.202 4. Accuracy of applied models is given below in table 4:
Table 4. Accuracy of Applied Models
From the results it is clearly shown that Gradient booster gives the maximum accurate results. The results are obtained currently using the Knime software but our future work is to develop an application so that the farmers can operate it easily.
473. [14] Raorane, A. A., and R. V. Kulkarni. "Application of DataMining tool to crop management system." Russian Journal of Agricultural and Socio-Economic Sciences 37, no. 1 (2015). [15] Rajak, Rohit Kumar, AnkitPawar, MitaleePendke, PoojaShinde, Suresh Rathod, and AvinashDevare. "Crop recommendation system to maximize crop yield using machine learning technique." International Research Journal of Engineering and Technology 4, no. 12 (2017): 950-953. [16] Savla, Anshal, Himtanaya Bhadada, Parul Dhawan, and Vatsa Joshi. "Application of machine learning techniques for yield prediction on delineated zones in precision agriculture." IJNCAA (2015): 48 [17] Son, Nguyen-Thanh, Chi-Farn Chen, Cheng-Ru Chen, Horng-Yuh Guo, Youg-Sing Cheng, ShuLing Chen, Huan-Sheng Lin, and Shih-Hsiang Chen. "Machine learning approaches for rice crop yield predictions using time-series satellite data in Taiwan." International Journal of Remote Sensing 41, no. 20 (2020): 7868-7888. [18] D. Sindhura, B. Navya Krishna, K. Sai Prasanna Lakshmi, B. Mallikarjun Rao, Dr. J Rajendra Prasad, Effects of Climate Changes on Agriculture International Journal of Advanced Research in Computer Science and Software Engineering,2016. [19] Van Klompenburg, Thomas, Ayalew Kassahun, and Cagatay Catal. "Crop yield prediction using machine learning: A systematic literature review." Computers and Electronics in Agriculture 177 (2020): 105709.