Abstract:

The economic crisis and the need for greater competitiveness require that private and public organizations significantly rationalize the processes of human resources selection, evaluation and development. These processes are based on five core activities: (1) analysis of production and management processes; (2) description of organizational profiles and roles; (3) investigation into the actual learning need of human resources; (4) design and delivery of training and development activities; and (5) evaluation of results. Beyond the peculiarities of the different production sectors, the common thread running through these five activities has always been the map of the competences needed by each professional profile to perform effectively in the organizational system. In effect, competences played a key role both in the definition of the criteria used to place human resources in the different functional areas of the company, and in the planning of training courses, paths for individual development and certifications.
Moreover, the transformation of the production mechanisms and the impact of new technologies call the traditional view of the concept of “competence” into question: the new technical knowledge is hard to relate to the existing frameworks, specialized skills keep on merging with management ones, and the increasing complexity of the new socio-economic scenarios requires a more adequate interpretation. Therefore, it is necessary to turn away from the old “linear” view of competences and to move towards the parallel and distributed nature of the cognitive and behavioral assets affecting performances in organizations.
In this paper, the authors propose a new view on the “competence” and present a reticular model of knowledge/skills analysis, based on a topological representation of the human capital through the distribution of the various profiles/roles within a two-dimensional map, with several levels of intensity. In particular, the basic idea is to identify the areas with a higher level of density as the key professional profiles in the organization, and then to place the different human resources according to these areas in the map. This classification must reflect the Euclidean distance between the focal points (the areas with the greatest density that correspond to the ideal best performers) and the points occupied by real people, through the assessment of patterns of observable behaviors that exhaustively define the identikit of each human resource in that particular organization.
From a practical point of view, the model proposes the implementation of a software module based on Kohonen Maps (Artificial Neural Networks trained by unsupervised learning). This module can gather information on human resources, process the different patterns that describe each person’s features and create a graphic representation of the distribution of cognitive and behavioral assets that are functional to the organizational processes. Therefore, it can efficiently support HR departments in several activities, like recruitment, definition of training plans (according to the distance between the position occupied by each person and the expected focal point), reorganization of the company and certification of new professional profiles.
A practical example shows the benefits of this new tool in case of human resources reallocation after relevant changes in the organizational structure.

Keywords:

Competences, Kohonen Maps, complexity, software module, unsupervised learning, topology, Human Resources.