In this paper, we aimed at presenting a methodology from which a good dissimilarity model can be obtained; to generate a true metric proximity matrix among paired objects and facilitate successful clustering procedure. In most cases, the property of the (dis)similarity model parameters is continuous in nature and as such requires integrations which, at times, are too complex to obtain in a closed form. This inability of the model integral to give a closed form solution is likely to make the realization of a good (dis)similarity model and its true metric proximity measure among paired objects not feasible. We presented a general and theoretically method for addressing the (dis)similarity model via a probability generating function which allows for contour integration and estimate the true metric measure of paired objects proximity as an area under a Fourier series curve. Dirichlet conditions were employed to ensure the convergence of the Fourier transformed (dis)similarity model to closed form. Result for the closed form (dis)similarity model and its proximity matrix on the measure of closeness of cases referred for psychiatric treatment as published by Marzillier and Field (2000) was established.