Li Niu

Cognition-Driven Decision Support for Business Intelligence

Cognition-driven decision support system (DSS) has been recognized as a paradigm in the research and development of business intelligence (BI). Cognitive decision support aims to help managers in their decision making from human cognitive aspects, such as thinking, sensing, understanding and predicting, and fully reuse their experience. Among these cognitive aspects, decision makers situation awareness (SA) and mental models are considered to be two important prerequisites for decision making, particularly in ill-structured and dynamic decision situations with uncertainties, time pressure and high personal stake. In todays business domain, decision making is becoming increasingly complex. To make a successful decision, managers SA about their business environments becomes a critical factor. This book presents theoretical models as well practical techniques of cognitiondriven DSS. It first introduces some important concepts of cognition orientation in decision making process and some techniques in related research areas including DSS, data warehouse and BI, offering readers a preliminary for moving forward in this book. It then proposes a cognition-driven decision process (CDDP) model which incorporates SA and experience (mental models) as its central components. The goal of the CDDP model is to facilitate cognitive decision support to managers on the basis of BI systems. It also presents relevant techniques developed to support the implementation of the CDDP model in a BI environment. Key issues addressed of a typical business decision cycle in the CDDP model include: natural language interface for a managers SA input, extraction of SA semantics, construction of data warehouse queries based on the mangers SA and experience, situation information retrieval from data warehouse, how the manager perceives situation information and update SA, how the managers SA leads to a final decision. Finally, a cognition-driven DSS, FACETS, and two illustrative applications of this system are discussed.

FACETS: A cognitive business intelligence system

A cognitive decision support system called FACETS was developed and evaluated based on the situation retrieval (SR) model. The aim of FACETS is to provide decision makers cognitive decision support in ill-structured decision situations. The design and development of FACETS includes novel concepts, models, algorithms and system architecture, such as ontology and experience representation, situation awareness parsing, data warehouse query construction and guided situation presentation. The experiments showed that FACETS is able to play a significant role in supporting ill-structured decision making through developing and enriching situation awareness.

A Situation Retrieval Model for Cognitive Decision Support in Digital Business Ecosystems

This paper presents a novel situation retrieval (SR) model for supporting cognition-driven decision processes in digital business ecosystems. Cognitive decision support in digital ecosystems is concerned with decision makers cognitive processes. This study aims to facilitate cognitive decision support to decision makers on the basis of current business intelligence (BI) platform. Underlying foundations of the SR model are two types of mental constructs: situation awareness (SA) and mental models of decision makers and the model of naturalistic decision making (NDM). These mental constructs and NDM are integrated into the BI application framework. Our experiments showed that the SR model was playing a nontrivial role to help decision makers develop enhanced SA and reuse their past experience to make better decisions.

An Exploratory Cognitive Business Intelligence System

An exploratory study of Web-based cognitive business intelligence systems (CBIS) is presented in this paper. The underpinning concepts and theories are situation awareness, mental model, and naturalistic decision making (NDM). The CBIS is an extension of the traditional business intelligence system with cognitive orientation. It focuses on developing, enriching, and utilizing the executives situation awareness, mental models, and other past experience during human-computer interaction, which drives the decision process to approach a naturalistic decision.

A Model of Cognition-Driven Decision Process for Business Intelligence

We proposed a cognition-driven decision process model for business intelligence. In this model, a managerpsilas situation awareness (SA) and mental models are developed and enriched for naturalistic decision making based on traditional business intelligence systems. Mental models are also used to supervise the process of situation information retrieval and presentation. The final decision-making process is based on recognition-primed decision model.

Cognition in Business Decision Support Systems

Cognition plays a key role in decision making for complex, ill-structured situations. Cognitive decision support is one of the major objectives in the design and development of DSS. In Chapter 3, we analyzed some important concepts and models in managerial cognition. Managers’ rich SA and mental models are two prerequisites for and likely lead to successful decisions according NDM theory. In this chapter, we will look at the role of managers’ cognition on decision-making processes from an information system (IS) perspective. The first two sections of this chapter discuss the basic characteristics of decision-making tasks in business domain, and how managers’ cognition can influence on business decision making. The last section summarizes some typical systems that take cognition as an important consideration.

Cognition-Driven Decision Processes

Human decision making is naturally a cognitive process conducted within a decision maker’s mind. However, the use of decision support systems incorporates IS artifacts into the human cognitive process. In this sense, decision making is an integral process comprised of the human cognitive process as well as IS functionality. We use the term cognition-driven decision process (CDDP) to denote such a decision making process. In this chapter, a conceptual framework is presented in Section 5.1, which shows high level ideas and essential characteristics of cognition-driven decision processes. Then the detail of the CDDP model is discussed in Section 5.2.

Natural Language Processing for Situation Awareness

We employ natural language processing (NLP) techniques to extract semantic information from SA descriptions. According to the CDDP model, managers can describe their SA using a natural language (English) and input SA descriptions into the system. This feature makes it easier for managers to interact with the system and develop their SA for decision making. SA descriptions are comprised of SA sentences. A syntactic parser is used to chunk down SA sentences into words and phrases, and extract syntactic relationships between words or phrases. Based on the results of the syntactic parser, instances, classes, SA triples and local contexts can be extracted or inferred. We refer to the whole process as SA parsing. A general English syntactic parser, Link Grammar Parser, is introduced in Section 7.1. SA parsing is, based on the information types defined in this book, discussed in Section 7.2. Section 7.3 is the overview of the process of SA parsing. The detailed sub-processes of SA parsing are presented in Sections 7.4, 7.5 and 7.6.

A Cognition-Driven Decision Support System: FACETS

Using the techniques discussed in Chapters 6, 7 and 8, we developed a prototype system, FACETS, as an implementation of the CDDP model. It allows a manager to describe his/her SA in the form of English. Based on the domain knowledge, FACETS parses the manager’s SA and constructs data warehouse queries. Queries are submitted to the data warehouse for retrieving relevant situation information. The retrieved situation information is presented to managers according to the navigation knowledge extracted from the manager’s experience. The goal of FACETS is to assist managers to develop and enrich their SA for decision making.

Evaluation of Algorithms and FACETS

This chapter reports on the experiments conducted to evaluate the techniques (algorithms) and system (FACETS) developed in this research. Based on a fictitious manufacture company, a data warehouse, an ontology and an experience base were developed and linked with FACETS. Students and professors were invited to the experiments. The FACETS and the major algorithms for SA parsing and query construction were evaluated based on the interaction between subjects and FACETS.