Measuring Variability and Factors Affecting the Agricultural Production: A Ridge Regression Approach

• Published in: Annals of data science 2023-04, Vol.10 (2), p.513-526
• Main Authors: Devi, Monika, Malik, D. P., Mehala, Vinay, Mishra, P.
• Format: Article
• Language: English
• Subjects: Agricultural production; Artificial Intelligence; Business and Management; Coefficient of variation; Cotton; Economics; Estimates; Finance; Insurance; Least squares method; Management; Regression models; Statistics for Business; Time series
• ISSN: 2198-5804; 2198-5812
• DOI: 10.1007/s40745-020-00274-0

The present investigation deals with the use of ridge regression in the presence of high degree multicollinearity, to determine the major factors influencing the agricultural production in Haryana state, India. Time series data for the period (1980–1981 to 2017–2018) on area, production, yield and different factors influencing the production of wheat, gram, paddy and cotton have been used in this study. Time series data was divided into four sub-periods of 10 year interval to measure the variability in area, production and yield of selected crops into different periods. Coefficient of variation around trend was also observed as the stability (or instability) index. Estimates for the factors affecting the agricultural production were observed by using the ordinary least squares technique and ridge regression both. A comparison is made and ridge estimates were found appropriate to measure the effect of different factors on agricultural production in the presence of multicollinearity. Total study period has been separated into training and testing parts and validity of the developed ridge regression model has been done on the basis of adjusted R 2 , percent deviation of forecast values and RMSE (%).