ESTIMATING THE EFFICIENCY OF MACHINE LEARNING IN FORECASTING HARVESTING TIME OF RICE
The study aims to measure the effectiveness of a system that used a machine-learning algorithm to predict the harvesting times of the rice crop. The study estimates the forecasting accuracy of the system using the opinions of experts and agriculture farmers. The system was designed to analyse the Moisture Content of the Grain, Ripe Grains Per Panicle, Number of Days After Sowing, Number of Days After Heading, Temperature at the time Heading and Surface Moisture of the soil. When the system is feed with before mentioned input, it generates output in the Likert scale of (Highly Not Recommended – Highly Recommended) using the mean score derived. For the study, 150 agriculture experts and 150 farmers practising rice cultivation for more than 5 years were considered as the samples. Further 24 fields not less than a hectare were considered for the study purpose. The opinion of agricultural experts and farmers were compared against the opinion generated by the system on the Likert scale. From the result obtained through analysis, it can be well perceived that there is no significant difference in an opinion posted by an expert, farmer with system generated output. Also, the system is generating output very lose to experienced farmed. Further, the standard deviation estimated is very least, which indicates the system is efficient enough in producing accurate opinion close enough with experienced farmer and experts. From the correlation value, it can be interpreted that there is a 79.1% relationship between opinions posted by an expert with system generated output. Further, there is a 72.8% relationship between opinion posted by experienced farmers with system generated output. Thereby, it is understood from the result that considering individual responses without considering the mean score, the system generated output is close to expert opinion. Furthermore, the estimated coefficient value indicates that the system generated output can be predicted using the opinion posted by Experts and Experienced Farmers. The regression equation is given; System Generated Opinion = 0.463 + (0.573×Expert) + (0.307×Farmers). From the regression plot, it can be additionally interpreted that most of the opinions posted by the Experts and Experienced farmers coincided with the opinion generated by the System using a machine learning algorithm.