Author(s)

A. Pagnoni, S. Parisi & D. Attorrese

Abstract

This paper suggests the use of data mining to design user-tailored on-line forms. By clustering the database of already submitted forms along relevant user attributes, it is possible to design user-specific forms that reflect the submission profile of each cluster of users. Such forms will contain only entries that are usually filled out by the users in the associated cluster; moreover, class-specific entries will be pre-compiled; all computations will be carried out automatically. Forms of this kind require less time to b fill out, and are less error-prone. The paper introduces this paperwork reduction strategy by means of an application for the Italian health and social security service. The unsupervised neural network used by our clustering algorithm is discussed in some detail. 1 Introduction Filling out forms—tax returns

Keywords