Treatment of missing and atypical values in the application of the Statistical Model of Impact Measuring in a study of 90 dairy farms in Pastaza province, Ecuador.

Main Article Content

E. O. Segura
Verena Torres

Abstract

The treatment of missing and atypical values is presented in the gathered information of a study of 90 dairy farms from Pastaza province, Ecuador, before applying the Statistical Model of Impact Measuring, to verify data reliability. The randomness joint test of Little was applied, according to the obtained value of Chi-square, not significant (P > 0.05), which confirms that data are missing completely at random, and makes possible the use of the attribution with regression analysis. Mahalanobis D2 and D distance values of all the analyzed farms were shown. These last were compared to the percentiles of Chi-square test: X2 (P < 0.05;20)= 34.2 distribution. The results showed the null hypothesis acceptance, so the farms were not considered as atypical cases. The used methods for the treatment of missing and atypical values allowed to obtain a complete data base of real values that will contribute to the application of the Statistical Model of Impact Measuring.

Key words: Statistical Model of Impact Measuring, joint test of randomness, Mahalanobis distance.

Article Details

How to Cite
Segura, E. O., & Torres, V. (2016). Treatment of missing and atypical values in the application of the Statistical Model of Impact Measuring in a study of 90 dairy farms in Pastaza province, Ecuador. Cuban Journal of Agricultural Science, 48(4). Retrieved from https://cjascience.com/index.php/CJAS/article/view/559
Section
Biomathematics