Recent standard ripeness classification for mango is via manual inspection by human naked eyes. However, the manual mango ripeness classification in agricultural setting has several drawbacks which need labor intensive, inconsistent, prone to error and it is also a time consuming process. Based on an extensive literature search, study to extract data patterns from mango images has never been conducted. Data pattern extraction or generally known as discretization, is one of data pre-processing method that stimulates classification. This paper presents the work on discretization that promotes classification process of mango (Mangifera Indica L.) dataset. Comparison between existing swarm-based discretization algorithms on mango dataset is studied throughout this paper in order to avoid inefficient manual effort and provide an improvement for future research in agricultural industry. The swarm-based discretization algorithm implemented on extracted features from mango images has reduced both discretization time and error rate concurrently. Hence, it generates good generalization of the data pattern to the extracted mango features. As a consequence, determining discretized data patterns from the extracted mango images may improve the entire classification process in terms of accuracy and learning time.
