Elsevier

Knowledge-Based Systems

Volume 85, September 2015, Pages 112-130
Knowledge-Based Systems

A new approach and insightful financial diagnoses for the IT industry based on a hybrid MADM model

https://doi.org/10.1016/j.knosys.2015.04.024Get rights and content

Abstract

Financial performance is vital for information technology (IT) companies to survive intense global competition. Because of the complexity in the business environment and the rapidly advancing technologies, companies lack specific guidance to understand the implicit relationship among crucial financial indicators for improving prospects in a contextual approach. To resolve the aforementioned concern, this study proposed a new approach by combining the variable consistency dominance-based rough set approach (VC-DRSA) with the decision-making trial and evaluation laboratory (DEMATEL) technique to explore the complex relationship among financial variables and improve future performances. In addition, a fuzzy inference system was devised on the basis of the findings of the VC-DRSA and DEMATEL technique to examine granulized knowledge and implications. A group of real IT companies listed on the Taiwan stock market were used as an empirical case to present the benefits of the new approach. The results generated a set of decision rules that can be used for forecasting future performance prospects and diagnosing the directional influences of crucial variables to gain insights; certain strong decision rules were further examined using fuzzy inference to verify the obtained implications. The findings contribute to the financial applications of decision-making science and computational intelligence in practice.

Introduction

Financial ratios are widely used for evaluating the competitiveness and worthiness of a company. This evaluation is often conducted by inspecting financial ratios of a specific industry or by comparing the current state of a company with its historical performance [35], and is termed as fundamental analysis (FA) [18]. By using FA, potential investors, shareholders, management teams, and external creditors may predict the financial performance (FP) of a company. Because of pressure from the capital market, it is crucial for the management teams of publicly listed companies to devise various plans (e.g., strategy planning, research and development roadmap, and financial planning) for improving FP. However, because of complex and rapidly changing business dynamics, obtaining a practical and understandable guidance for achieving this goal remains a challenge. Thus, considering the rapid advancement in technological innovation and the intense competition in the IT industry, this study focused on the FP analyses of IT companies.

In conventional studies, researchers primarily relied on statistical models to investigate the relationship between financial ratios and the subsequent change in the performance of businesses [1], [20], [28], [33], predominantly the FP (e.g., earnings per share growth rate or stock returns). The main difference in these studies was the included variables and explained performance indicators. Although this approach is widely used in financial studies, the unrealistic assumptions of regression models (e.g., independent relationship among considered variables and normal distribution of errors) might yield unpersuasive results [24]. Furthermore, the regressions primarily represent average results, which are insufficient to guide a decision maker (DM) [44]. Therefore, the complexity of multiple dimensions and criteria enticed researchers from other fields to resolve performance prediction problems, such as multiple attribute decision making (MADM) [7], [19] and computational intelligence [8], [18], [34].

Although the performance prediction problems have gained attention in various research fields, most studies have used a subjective approach to collect the knowledge of domain experts for modeling [19], [36], [41], [51] or the data mining approach to explore implicit relationships among large datasets [9], [12], [34]. An integrated model that can be used to retrieve useful knowledge from the two aforementioned approaches requires further exploration. Therefore, this study proposed a new approach for determining the FP by integrating the computational intelligence model and the knowledge of experts to solve the FP prediction problem. This study initially inducted decision rules from a group of real IT companies by using an extended rough set approach (RSA), and then used the retrieved core attributes to collect the knowledge of experts, thus guiding a company to analyze its FP by illustrating the influences of certain criteria or dimensions in a specific context. In addition, to verify the obtained knowledge, a fuzzy inference system (FIS)—based on strong decision rules and the decision-making trial and evaluation laboratory (DEMATEL) analysis—is devised to examine the reasoning logics (i.e., knowledge or implications).

To understand the core attributes that may enable the prediction of subsequent performance change in the IT industry, this study proposed a new approach for attaining an insightful analysis. This study attempted to obtain comprehensible decision rules by considering the relative importance and directional influences of various criteria (i.e., financial attributes or ratios). The proposed model provided a diagnosing tool and a process that might determine the influences shaping future prospects. The insights acquired during the process should be meaningful to support strategy formulation, planning, and decision making, which could not be achieved through statistical analyses. The implications are expected to provide constructive meanings both in academia and practice.

The remainder of this paper is structured as follows: Section 2 briefly introduces FA and certain MADM methods related to this study, Section 3 describes the proposed model, Section 4 presents an empirical study of real IT companies in Taiwan, and Section 5 describes the empirical results. Finally, Section 6 concludes the study and discusses future research directions.

Section snippets

Preliminaries

This section briefly reviews the methods and techniques used in this study, including the variable consistency dominance-based rough set approach (VC-DRSA) and the hybrid MADM methods (including the DEMATEL technique and fuzzy inference).

Hybrid MADM model based on VC-DRSA and DEMATEL techniques

This section presents the proposed new hybrid model for forecasting the FP of IT companies. The model comprises three parts, namely the VC-DRSA (to induct decision rules and identify crucial financial variables for analysis), DEMATEL analysis (to acquire the directional influences among core criteria and dimensions), and FIS (to validate the obtained knowledge by experts).

Empirical case of the information technology industry in Taiwan

A group of IT companies in Taiwan were analyzed for exploring the key indicators used for forecasting the FP. The opinions of domain experts were collected using questionnaires. The proposed model comprises three parts; a diagram of the involved steps is shown in Fig. 2.

Discussions and implications

The VC-DRSA model (CL = 0.95) generated 42 decision rules to classify the subsequent FP of the IT companies, and achieved more than 85% accuracy of approximation for both the training set and testing set (Table 3). In addition, the attributes obtained were analyzed using the DEMATEL technique to collect the directional influences among the CORE. The combination of decision rules and directional influences among the criteria may thus provide an insightful guidance for IT companies to improve.

Conclusion and remarks

This study proposed an integrated approach that utilizes the advantages of two approaches for improving the FP of the IT industry: the VC-DRSA and DEMATEL technique; in addition, the integrated DFG (infusing decision rules and the INRM) provided DMs additional insights on improvement planning. Furthermore, this study examined the implications in the DFGs (Fig. 5a, Fig. 5b) by adopting fuzzy inference, which processed the granulized concepts and supported the findings. Fig. 7 depicts a diagram

References (53)

  • J.-J. Liou et al.

    A dominance-based rough set approach to customer behavior in the airline market

    Inform. Sci.

    (2010)
  • J.-J. Liou et al.

    Using decision rules to help airlines cut costs without sacrificing essential service quality

    Eur. J. Oper. Res.

    (2010)
  • E.H. Mamdani et al.

    An experiment in linguistic synthesis with a fuzzy logic controller

    Int. J. Man–Mach. Stud.

    (1975)
  • Y.-P. Ou Yang et al.

    A VIKOR technique based on DEMATEL and ANP for information security risk control assessment

    Inform. Sci.

    (2013)
  • V. Ravi et al.

    Soft computing system for bank performance prediction

    Appl. Soft Comput.

    (2008)
  • S. Richardson et al.

    Accounting anomalies and fundamental analysis: a review of recent research advances

    J. Account. Econ.

    (2010)
  • A. Sala et al.

    Perspectives of fuzzy systems and control

    Fuzzy Sets Syst.

    (2005)
  • K.-Y. Shen et al.

    Combining VIKOR-DANP model for glamor stock selection and stock performance improvement

    Knowl.-Based Syst.

    (2014)
  • G.-H. Tzeng et al.

    Evaluating intertwined effects in e-learning programs: a novel hybrid MCDM model based on factor analysis and DEMATEL

    Expert Syst. Appl.

    (2007)
  • W.-W. Wu

    Choosing knowledge management strategies by using a combined ANP and DEMATEL approach

    Expert Syst. Appl.

    (2008)
  • J.-L. Yang et al.

    An integrated MCDM technique combined with DEMATEL for a novel cluster-weighted with ANP method

    Expert Syst. Appl.

    (2011)
  • L.A. Zadeh

    Toward a theory of fuzzy information granulation and its centrality in human reasoning and fuzzy logic

    Fuzzy Sets Syst.

    (1997)
  • J.S. Abarbanell et al.

    Fundamental analysis, future earnings, and stock prices

    J. Account. Res.

    (1997)
  • J. Błaszczyński et al.

    jMAF-dominance-based rough set data analysis framework

  • J. Błaszczyński et al.

    Monotonic variable consistency rough set approaches

  • M. Doumpos et al.

    Computational intelligence techniques for multicriteria decision aiding: an overview

    Multicriteria Decis. Artif. Intell.: Links, Theory Appl.

    (2013)
  • Cited by (0)

    View full text