Abstract:
Every aspect of this digital world is making a huge difference due to the impact of
the ITsector. As a country that has an emerging economy, Sri Lanka’s agricultural
sector needs more assistant for improvement. Forecasting the prices of the
agricultural products is in acritical situation due to lack of technology adoption for
that. Price forecasting helps farmers and the government to make an effective
decision. In recent years, crop prices have been changed dramatically due to
unpredictable climate change, natural disasters, and other problems. Farmers are
unaware of these uncertainties and they incur huge losses. In addition, from the Sri
Lankan point of view, there is no evidence that crop prices were predicted using
machine learning. The main objective of this research is to predict crop prices and
fill in the gaps in the literature using machine learning methods. Data mining is
emerging as an important field of research in agricultural crop price analysis. In
this study, researchers have discussed methods of forecasting vegetable prices,
which are aimed at farmers, government, consumers, and other stakeholders
focusing on profitable vegetable cultivation in Sri Lanka. The population of the
present study is all the vegetables available in the vegetable market in Sri Lanka.
Among them, Beans, Brinjal, Carrot, and Pumpkin were selected as the sample for
this study by employing the systematic sampling method. The analysis was
performed using five different classifiers and the results were evaluated using the
mean absolute error, root-mean-square error, relative absolute error, and root relative square error. The predicted data were compared with the actual data and
this model created a significant level of accuracy. Artificial neural network
classification predicted the best results of the model. Vegetable prices have high
nonlinear and high noise-like properties. Therefore, it is difficult to predict
vegetable prices.