Linear Discriminant Analysis of Multiple Groups in Rural Settlements of Akwa Ibom State, Nigeria

This study examined the levels of stock of social infrastructure and the spatial pattern of development in rural areas of Akwa Ibom State, Nigeria. Empirical and theoretical approaches were employed in the investigation and data on 21 social indicator variables/surrogates were collected from 50 villages in the State using questionnaire and field observation as research tools. An index of social infrastructure stock was evolved and hierarchical cluster analysis statistics was applied on the stock of social infrastructure in order to group the communities on the basis of social infrastructure profiles. The single linkage cluster analysis was employed to illustrate the linear combination of the communities in rural areas that were found to fall into low (Group 1), fair (Group 2), moderate (Group 3) and high (Group 4) performance patterns of social infrastructure stock. The result shows that the study area is characterised by many vulnerable communities that are very weak in stock of social infrastructure. The multiple linear discriminant Analysis (MLDA) technique was used to assess the optimality of earlier groupings of settlements in the study area. The result showed that MLDA correctly classified 97.6 per cent of the settlements. The technique correctly classified most of the Group one settlements with a few misclassifications but correctly classified all the remaining groups of settlements without any misclassification. In addition, health infrastructure was identified as the single most important independent variable that discriminated the four groups of settlements obtained earlier, thus highlighting its contribution to improving the social infrastructure in the study area.