Abstract
Large amounts of collected and stored data require specialized processing skills. However, a human-computer interaction is crude and far from natural. Users are forced to interact with databases using languages understood by machines. A simple way to learn more about phenomena represented by data can be done via representing the data as human-perceived concepts and enabling a human-like interpretation of it. The users should be able to use linguistic terms – for example LARGE, SLOW, MOST – as their representations of concepts in order to gain a better understanding of data. Yet, other issues arise: how to enter definitions of such terms, how to incorporate individual's understanding of their meanings, how to ensure their proper interpretation, and of course, how to do all this in an easy and simple way.
In this paper, we present and describe an iPad-based software that enables an easy procedure of defining linguistic terms – such as LARGE, MEDIUM, SMALL – and linguistic qualifiers – ALL, MOSTLY – that are suitable for data analysis purposes. We state, that linguistic terms and qualifiers represented as fuzzy sets embody human defined concepts and allow users to better understand data. This understating is achieved via ‘mapping’ data into models built with definitions of imprecise concepts familiar to users. Further, fuzzy-based definitions of concepts and fuzzy operations on them facilitate a human-like analysis of data as an important aspect of data science. The main contribution is a tablet application – called Tablet input of Fuzzy Sets (TiFS) – that allows users to define terms in a simple way via drawing their ‘shapes’ using fingers. We provide a detailed description of the developed iPad application for that purpose and show its benefits when used in defining linguistic terms for data summarization processes.