Model Development in Predicting Seaweed Production Using Data Mining Techniques

Abstract

Production trends nowadays can be predicted by identifying hidden patterns in variables and factors underlying the production processes. In the industry sector, data mining is a valuable data analysis method that can help forecast variabilities in the different industrial processes, particularly, the production area. Useful knowledge can be analyzed from data in databases that can be very helpful in identifying specific factors that improve the quantity and quality of products. Thus, this study utilized data from an agency database that was analyzed using data mining techniques to develop a model on seaweed production. Seaweed production is continuously increasing its share in the market but is also faced with problems and constraints. This study determined significant variables that can predict high seaweed production and compared classification accuracies of Naïve Bayes, J48, CART, logistic regression and CHAID algorithms. A prediction model on seaweed production is generated that can benefit the seaweed industry sector in designing interventions to increase its production. Validation of the result through other data mining techniques is recommended by this study.