Discriminant analysis for prediction and classification of farmers based on adoption of drought coping mechanisms

Abstract

SUMMARY :The study was carried out to develop the classificatory statistical model to predict and classify the farmers into adopters and non-adopters in Kolar district of Karnataka for the year 2013. Linear discriminant analysiswas carried out by considering the various socio-economic characteristics of farmers as predictors and adoption behaviour of the farmers as response variable in order to assess the factors influencing on adoption of drought coping mechanisms. The result shows that the Box's M test is 161.3 with their F approximation 1.83 is non-significant (0.19) at 5% level of significance, Eigen value (2.51) of the first function explains 100% of variations in the data which is potential enough to classifying the groups, wilk's lambda associated with the function (=0.28) is transforms to a chi square of 140.82 with 12 DF, which is statistically significant and the following variables such asFarm Size (0.552), Extension Visits (0.574), Crop Diversification (0.321) and Crop Insurance (0.368) are relatively more important and positively influencing on discrimination of farmers group. Whereas the variable like Age (-0.516) negatively influencing on discrimination of adopters and non-adopters.

Determinants of climate change adaptation strategies in South India: Empirical evidence

“…This implies that the farmers who enjoy better access to climate change information (i.e., seasonal or mid-term forecasting) made betterinformed adaptation decisions. These findings were similar to the findings from various studies (Belay et al, 2017;Halagundegowda et al, 2017;Adeagbo et al, 2021;Tshikororo et al, 2021). As expected by the researchers, education and livestock rearing had a positive association across all climate change adaptation strategies.…”

“…This indicates the product-moment correlations between the discriminating variables and discriminant function. Factor loadings of ≥0.30 were used as the cutoff between important and less important variables, that is, if the structure coefficient was ≥0.30, it was considered meaningful (Duong et al, 2017;Halagundegowda et al, 2017;Kumari et al, 2017). The findings indicated that the structure coefficients with the highest relationship to function 1 were OFFI (0.846), FI (0.789), and FE (0.730).…”

Adaptation of farmers to climate-change and determinants of their adaptation decisions in South India

“…This implies that the farmers who enjoy better access to climate change information (i.e., seasonal or midterm forecasting) made better informed adaptation decisions. These ndings are similar to the ndings from various studies(Adeagbo, 2021;Halagundegowda, 2017;Tshikororo, 2021). As expected by the researchers, EDU and livestock rearing had a positive association across all climate-change adaptation strategies.…”

“…A high canonical correlation coe cient (0.948) indicates a function that discriminates well between two categories of dependent variable and also infers no overlapping among them. Squaring the canonical correlation suggested that 89.8 per cent of the variation in the grouping variable was explained (Halagundegowda, 2017;Nguyen, 2017). As shown through Table 4, unexplained error is 10.2 per cent (Wilks' λ: 0.102).…”

“…It indicates the product moment correlations between the discriminating variables and discriminant function. Factor loadings ≥0.30 is used as the cut-off between important and less important variable ie., if the structure coe cient is ≥0.30, it is considered meaningful (Halagundegowda, 2017;Kumari, 2017;Nguyen, 2017). The ndings indicated that the structure coe cients with the highest relationship to function 1 were OFFI (0.846), AI (0.789) and FE (0.730).…”