Human-Computer Interaction for Knowledge Discovery for Management

Introduction

The relevance of data-intensive sciences is growing as a result of the exponential development in data quantity and complexity, rising processing power, and the availability of new computer technologies (Kousez et al. 2009). The enormous, complicated, and sometimes illogically or completely unstructured data is one of the major problems of the networked 21st century (Holzinger et al. 2013). New, effective, and user-friendly data processing technologies are needed to handle the data's growing volume. Traditional techniques for data interpretation frequently fall short of end users' rising expectations. It's interesting that several sophisticated computational tools have recently been created by independent groups with various philosophies: Researchers in data mining and machine learning frequently have faith in the ability of their statistical techniques to find important patterns, often automatically and without human intervention. However, as end user comprehension and control are reduced, the risks of modelling artefacts increase (Shneiderman et al., 2001, 2002).

As computer systems (Von-Neuman machines) lack the “plastic” components that make up the nervous system that are fundamental to human thinking, they run the danger of producing outcomes that are subpar. Peter Naur asserts that a unique, non-digital method is necessary to comprehend human thought, citing Synapse-State theory as one example (Naur et al. 2008).

As Herbert Simon recognized 40 years ago, an abundance of data leads to a poverty of concentration, and it is required to distribute that attention effectively among the overwhelming variety of information that may absorb it. Moreover, this data avalanche has been expedited by mobile, ubiquitous computers, ubiquitous sensors, and cheap cost storage. In order to deal with this predicament and the expanding data deluge, it is important to strive towards ensuring efficient human control over powerful machine intelligence through the integration of methods involving machine learning and visual analytics. To aid human comprehension and decision-making, cutting-edge computational and user-centered approaches must be used.

A creative energies of techniques, methodologies, and strategies from two fields—HCI, with its significance to intelligence, and KDD, with its focus on cognitive computing—provides the ideal framework for addressing these challenges. KDD aims to support human intellect with machine intelligence by uncovering novel, hitherto undiscovered insights within the sea of data. The main contribution of HCI-KDD is to enable end users to discover and recognize previously unknown and possibly relevant and useful information, in accordance with the principle that “science is to test hypotheses, engineering is to put these notions into business” (Holzinger et al. 2011). It is the process of identifying new, trustworthy, and maybe useful data patterns with the goal of understanding these patterns for decision-making.

This chapter provides a thorough overview and synopsis of the key fundamental subjects, such as Human Computer Interaction (HCI) and Knowledge Discovery (KDD). The interaction of HCI with other areas is then discussed, followed by a review of the KDD process and all of its stages, including Data Selection, Cleansing, etc., with a focus on the Data Mining stage. The peculiarity of the HCI-KDD combination is then briefly explained. It also offers potential paths for more research in its conclusion. The major benefit of this chapter is that it might serve as an excellent introduction for people who are new to the subject.