Prediction of Feed Quantity Using Multiple Linear Regression Algorithm for Clarias Farming

One of the prevalent fish commodities in the community is clarias which is marked by high demand. High demand makes clarias farmers need an effective aquaculture strategy to assemble the consumption needs of clarias in the community. One strategy that needs to be noticed is the determination of the appropriate feed size so that the feeding is maximized. Currently, farmers have difficulty in determining the right size of feed, as a result, many uneaten feed remains are found. Therefore, an analysis of determining the size of the feed is needed that is adjusted to the width of the clarias's mouth so that feeding can carry out optimally. This study analyzed determining feed size in clarias using the Multiple Linear Regression algorithm. This study uses 150 data, namely training data totaling 105 and testing data totaling 45. The result of the coefficient of determination obtained is 81.40 and has an intercept value of -0.094. Measurement of the performance of the Multiple Linear Regression algorithm equation using Mean Absolute Percentage Error shows a result of 21%. From the results of the analysis above, it concluded that the determination of feed size affects the optimal growth of clarias.