THE NEED FOR MULTIDIMENSIONAL POVERTY ANALYSIS APPROACH Multidimensional poverty analysis aimed at gaining an understanding of the local

situation in its wider context, including social, economic, physical and financial aspects, unlike the other economic poverty analysis which tries to explain the situation on the context of economics direction leaving the other face of poverty. Consequently that approach makes poverty analysis a sided approach

A useful literature was provided from the contributions of the following on multidimensional poverty. (Praag and Carbonell, 2005; Kamanou, 2005; Anup, 2010; Foster, Greer and Thorbecke, 2010; Sabina and Foster, 2011; Rodrogo, 2011; Alkaire and Foster, 2011a; Alkire and Foster 2011b; Hatzimasoura and Bennet, 2011; Gaston 2012). Their effort to justify the need for multidimensional poverty is clear. Most of them observed that poverty analysis is tilted on economic perspective while leaving other aspect. This is quite a complete movement away from traditional one-dimensional to multidimensional poverty measurement particularly with respect to the identification step.

There were consensus that poverty depends on many variables not just income, expenditure or wealth. Longevity, low education poor health status, insecurity, low self confidence or powerlessness, absence of right such as freedom are the most obvious variable. Most of them if not all are ordinal variables in nature they are rank based on specific criteria. Also an Individual cannot be said is poor completely or not since He/ she can be poor in some dimension and not poor in another dimension. Hence it is good to measure poverty as degree or intensity of being in poverty rather than classifying them into poor and non poor. Even the most widely used model for income poverty analysis, the FGT (Foster Greer and

views. Obviously. But must transform to reflect the way analysis can be possible. This transformation are done by Composite or latent variable measurement. Recent advances in computer solve these problem now many classification. Hence require a new insight into how to capture the perspective of the affected the way they experience it directly. reduction and pattern. This can be done through multidimensional approach to capture data. totally random difficult and inappropriate in most inferential data analysis.Thorbike) which was developed 25 years ago now change perspective try to capture multidimensional poverty using ordinal variables. This process is one of many methods of solving the voluminous nature of poverty indicators which were thought to be a problem before. principles and feelings to the entire situation. The major problems with the ordinal and categorical variable is that most such data at first look researchers feel embarrassed. nature and intensities of their various deprivation and deficit that is entirely different from that of the analyst studying a particular indicator. Notwithstanding the fact that the methodology of self perception based approaches to poverty identification and analysis is not without its own problems the relevance of such subjective information cannot be overlooked. Recent work tries to use the multidimensional indicators but reduced and transformed them in a form more liable to further analysis. especially absolute income or food poverty line. forms. The concept of “Attitude” “Belief” and “Opinion” had been assessed in many behavioral research using either ordinal or Categorical measurement scales. it is a clear fact those who are experiencing poverty usually have a perspective on the sources. Generally the subjectivity of poverty depends on the individual. These insights remain silent and invisible. . relationship detection software are available to which uses both ordinal and categorical variables objectively.

discriminant analysis. This makes multidimensional poverty analysis intricate task. Factor. variable interacting between themselves. component. the use of multivariate statistical technique in multidimensional poverty analysis increased significantly. None the less. cluster analysis techniques were suitable for analyzing the patterns of complex. and forms the core of many disciplines. grouped. The pattern recognition is the process of finding general relations in a set of data. most of the problems experience using regression. multidimensional scaling. therefore for ease of investigation some statistical applications had to be imploded. Measurement of relationship and pattern detection in multidimensional poverty is done with multivariate analysis technique which comprise of factor analysis. classified. cluster analysis conjoint analysis.Multidimensional poverty is complex interrelated problems and mostly two way dependency relationships exist. multidimensional relationship encountered in research. Karl . These techniques prove beneficial in their application and solve. and correspondence analysis. principal component. or segregate into a smaller set of factors or components. examining relations are the common approach to multidimensional poverty analysis. Pattern recognition. classification. They can be utilized to examine the underlying pattern or relationship for large number of variables and to determine whether information can be condense or summarized. from neural networks to so-called syntactical pattern recognition. Generally in multidimensional poverty analysis number of variables exhibit interrelationship between themselves hence the need for knowledge of the structure patterns of the variables.

identification and simultaneous equation bias and the effect of aggregation problem common to many macro economics problem. 2006. The techniques of cluster analysis have been . Bengt and Tihomir. The ultimate strategic options based to identifying groups within the populations such as segmentation and classification would not be possible without an objective methodology. Abdeljaouad and Paolo (2012) in Morocco. Koutsyiannis. Manly. 2001. and converting them to metric variables usable for analysis. their level or even their behaviors. pattern recognition and relationship exploration. The approach helps to avoid multicolinearity problem. But multiple correspondence analysis approach was used to measure multidimensional poverty by Herve. Although Rencher (2002). 2001. The method is comprehensive embracing every aspect of human life at the same time transforming and serving the objective of data reduction. 2007. all suggested principal component or factor analysis for data reduction without losing the essential features of the indicators as well as capturing ordinal. or individuals suffering from poverty. Assellin (2009) in Senegal. 2005 Ken. Cluster analysis is the best approach to perform such.and Irini. all believe that principal component and factor analysis are good in data reduction for both metric or non metric data In Multidimensional poverty it is common to encountered situations that can best resolve by defining groups of homogenous variables. categorical. 2011. The multiple correspondence analyses were principally used for data that were categorical or nominal unlike the factorial and principal component method which uses both. Landau(2004). Hair (2010).

geography. remote sensing. The attempt is to maximize the homogeneity of the objects within the cluster while also maximizing the heterogeneity between the clusters. taxonomy of plant and animal.extensively applied to data in many fields. soil classification and engineering. psychiatry. The method use in searching for natural structure among the observation based on multivariate profile. Cluster analysis group individuals or objects into clusters so that objects in the same cluster are more similar to one another than they are to objects in other cluster. sociology. archaeology. geology. such as medicine. economics. criminology. . market research. Therefore this work will used the approach in searching for groups and classes of multidimensional poverty in the area as one of the analytical tools. anthropology.