Tags:Classification, Data mining, Naive Bayes, Rule JRIP and Tree LMT
Abstract:
Data mining is a classification technique that can be used to handle large volumes of data. Hence, data mining has evolved as an excellent solution for large agricul-tural datasets. This is partly because it can predict categorical class labels, classify data based on training set and class labels, and it can also evaluate new data. In agricultural production, farmers and agribusiness representatives need to make daily decisions. However, accurate yield estimate of the various crops related to the planning is a critical issue for agricultural plannings. Data mining technique is therefore required for achieving realistic and effective outcomes. The aim of this study is to classify different data features and implement various algorithms as it relates to agricultural big data. Additionally, a given dataset is preprocessed to en-sure that relevant data is present in all datasets. Algorithms such as Rule JRIP, Tree LBT, and Naive Bayes are implemented. Then, the Mean Absolute Error (MAE) and Relative Absolute Error (RAE) were compared, and the performance error of the resulting classification algorithm is performed on each dataset. The overall results indicates that JRIP has the highest efficiency with a value of 96%. This is followed by Naive Bayes which has 84% efficiency, whereas tree LMT has 78% efficiency. The result of this study can help to advance current research, as well as benefit future research in the agricultural sector.
Analysis of Data Mining Algorithms for Predicting Rainfall, Crop, and Pesticide Types on Agricultural Datasets