Application of Variable Selection Methods in Pineapple Production Modeling: A Case Study of Manipur, India

Pineapple (Ananas comosus) is one of the most significant tropical and subtropical fruits in the Bromeliaceae family, accounting for over one-third of global production. It's one of the most prominent commercial fruit crops in the world. In north-eastern India, Manipur was one of the prominent states that produced pineapples. The pineapple varieties cultivated in Manipur were Kew and Queen. Economically, the fruit has also become the backbone of a sizeable section of farmers who have been cultivating it as their major source of income. Therefore, in order for the government to develop plan that would sustain or enhance output, it was necessary to identify the factors that contribute to pineapple production. The main objective of this study was to identify the variables influencing the production of pineapple in Manipur. Primary data were collected for the practical application of the model. The sample size is decided at 450 for this consideration. The sampling design for this survey was stratified three stage sampling schemes of unequal size. Multiple linear regression model was fitted for the study. The stepwise regression estimation procedures and testing of relevant hypothesis specifically the Student’s t test and Snedecor’s F tests were found useful in regression analysis and analysis of variance that followed. The value of R2 for the model obtained at the 5th step i.e., 0.933 indicates that out of the total variation 93.3% of the variation was explain by the independent variables included in the model. The coefficients obtained from the model namely Xi4 (total area), Xi11 (average income), Xi10 (average expenditure), Xi9 (number of times of weeding per year) and Xi8 (amount of credit in rupees) were all found significant.