Symbols-Meaning-Value (SMV) space as a basis for a conceptual model of data science
Introduction
The theory of three-way decision (3WD) is supported by three pillars: the philosophy of thinking in threes, the methodology of problem-solving in threes, and the mechanism of computing (i.e., information-processing) in threes [59]. In everyday life, we are constantly making all types of decisions, ranging from small and trivial to large and complex. We are making decisions consciously, subconsciously, or unconsciously. To build a theory of three-way decision, I use the word ‘decision’ as an umbrella term to cover a broad category of activities, processes, and methods that help us in making decisions or that we actually use to make decisions. Under this interpretation, the theory of three-way decision embraces various three-way constructs, ideas, and models, including three-way thinking, three-way problem-solving, and three-way computing. For the current state and development of the theory, one may read the papers by Yang and Li [52], Wei et al. [51], and Yao [53] on networks analysis and bibliometrics analysis of research in three-way decision.
The philosophy of thinking in threes or triadic thinking appears across different fields. Triadic and tripartite understanding, formulation, and description of natural and human-made things and systems are indeed a common practice. For example, Pogliani et al. [38], [39] used many examples to show that we humans in general and scientists in specific have an intriguing preference for a ternary patterned reality. They suggested that “the model based on a ternary pattern ended up assuming an archetypical character, which is quite intriguing, and deserves to be deepened.” In selecting the most important ideas in the history of thought and invention, Watson [50] commented that “numerous figures in the past have viewed intellectual history as a tripartite system – organised around three grand ideas, ages or principles.” In a study of marketing eschatology, Brown [5] pointed out that, in general, “thinking in threes is a very common occurrence” and “trinitarian inclination is made manifest in many ways, shapes and forms,” and, in specific, “trinitarianism is widespread in the theory and practice of marketing.” For additional sources of information and examples of thinking in threes, a reader may consult my paper on the geometry of three-way decision [59].
In this paper, I explore what three-way decision can offer to a relatively new discipline known as data science (DS) [8], [11], [21], [37], [64], [65]. The trifold objectives are to examine the uses of the principles of triadic thinking in data science, to introduce a concept of the symbols-meaning-value (SMV) space or the SMV triad, and to construct a conceptual model of data science based on the SMV space. The SMV triad represents the three important aspects of data, namely, the form, the content, and the utility. Compared with existing studies, I will focus on a high-level conception of data science rather than investigating various concrete and particular theories, methods/algorithms, and systems/applications. With the SMV triad, I suggest a definition of data science as a field of scientific investigation of data with a focus on data-processing at the symbols level, knowledge-seeking at the meaning level, and wise decision-making at the value level. I metaphorically describe these three main tasks as seeing, knowing, and doing. In more technical terms, the three categories of tasks are 1) data collection, storage, retrieval, etc., 2) data analysis, data mining, machine learning, knowledge discovery, etc., and 3) data-grounded decision-making, data-grounded action, data-driven applications, etc. The main benefit of this unique and different kind of exposition is that the conceptual model may be helpful in understanding and guiding how we perceive, how we think, and what we do in data science.
The rest of the paper is organized as follows. Section 2 presents three examples of triadic thinking in data science and motivates the present study. Section 3 examines three interpretations of a triad and sets up the ground work. Section 4 introduces the concept of the symbols-meaning-value (SMV) space and discusses three perspectives on the SMV space from the viewpoints of information science and management science, cognitive science, and computer science. Section 5 articulates a conceptual understanding of data science based on three structural interpretations of the SMV space. Section 6 discusses practical values and implications of the conceptual model. Finally, Section 7 gives some concluding remarks.
Section snippets
Three examples of triadic thinking in data science
Given the abundant appearances of thinking in threes in different disciplines and fields, one would expect to see the same pattern in data science. I will use three examples to show that this is indeed the case.
The first example is a 3D conception of data management given by Laney [24], which is widely referred to as the three V's of big data, consisting of data volume, velocity, and variety. While these three V's capture some of the main features of big data in the aspect of the form of the
Three interpretations of a triad and three constructive methods
A triad consists of three things and the three things may be unstructured, partially structured, or structured in various ways [59]. The generality, flexibility, and universality of a triad makes it suitable to serve as a basic notion of three-way decision. By interpreting a triad in different contexts, it is possible to obtain an array of models, modes, and tools of three-way decision. Among the many triadic structures and patterns [59], I briefly discuss the three in Figs. 1(a)-(c) and the
Symbols-meaning-value (SMV) space
Based on Weaver's trilevel classification of communication problems [43], I introduce the concept of the symbols-meaning-value (SMV) space. The SMV space may serve as a basis of a conceptual framework for describing, understanding, and structuring what and how we perceive, what and how we know, and what and how we do. By virtue of its simplicity and flexibility, the framework, hopefully, is applicable to many disciplines, various fields, and different domains. To support this claim, I discuss,
SMV conceptual model of data science
Based on the ideas of the SMV space and its various interpretations, I present a general conceptual understanding of data science. A specific interpretation of the SMV space as the DKW (data-knowledge-wisdom) hierarchy seems to be suitable for describing data science. I introduce two variations of the DKW hierarchy for considering other types of relationship between the levels of the DKW hierarchy. One is the DKW stair in Fig. 3(b) and the other is the DKW circle in Fig. 3(c). Together with the
The need for the SMV model and its values and implications
Given the description of the SMV model of data science in the last section, many questions arise naturally with respect to the need for such a model and its values and implications. In fact, the reviewers of the paper suggested three such questions: Why do we need yet another model of data science? What does the SMV model offer that is not as easily understandable within the classical perspectives on data science? What are the practical implications of the SMV model? Answers to these questions
Conclusion
In this paper, I explained why, what, and how the theory of three-way decision contributes to data science. I introduced the concept of the SMV (symbols-meaning-value) space and used it as a basis for developing a conceptual model of data science. It gives a 3×3 conception with respect to the combinations of three levels and three issues. At the symbols level, data science focuses on the form of data, treats data as a resource [25], [36], and has the objective to make data available. At the
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This paper was based on a keynote speech delivered at The 7th International Conference on Big Data and Information Analytics (BigDIA 2021). I would like to express my thanks to Professor Guoyin Wang for his encouragement and support. I am grateful to reviewers for their encouraging comments and suggestions, which resulted in the introduction of Section 6. This work was partially supported by a Discovery Grant from NSERC, Canada.
References (65)
- et al.
The three-way-in and three-way-out framework to treat and exploit ambiguity in data
Int. J. Approx. Reason.
(2020) Three-way data analytics: preparing and analyzing data in threes
Inf. Sci.
(2021)The knowledge level
Artif. Intell.
(1982)The two sides of the theory of rough sets
Knowl.-Based Syst.
(2015)Three-way decision and granular computing
Int. J. Approx. Reason.
(2018)- et al.
Towards data science
Data Sci. J.
(2015) From data to wisdom
J. Appl. Syst. Anal.
(1989)Planning and Control: A Framework for Analysis
(1965)Seeing, Knowing, and Doing, a Perceptualist Account
(2020)What data scientists really do, according to 35 data scientists
Harv. Bus. Rev.
(2018)
Trinitarianism, the eternal evangel and the three eras of schema
Three-way decision for handling uncertainty in machine learning: a narrative review
Data science: a comprehensive overview
ACM Comput. Surv.
Leading with decision-driven data analytics
MIT Sloan Manag. Rev.
Discourse on the method of rightly conducting one's reason and seeking truth in the sciences
Data science and prediction
Commun. ACM
Toward foundations for data science and analytics: a knowledge framework for professional standards
The knowledge pyramid: a critique of the DIKW hierarchy
J. Inf. Sci.
Research Methods: A Process of Inquiry
Data, DIKW, big data and data science
Proc. Comput. Sci.
Some comments on big data and data science
Ann. Data Sci.
Can robots possess knowledge? Rethinking the DIK(W) pyramid through the lens of employees of an automotive factory
Humanit. Soc. Sci. Commun.
The trilogy of mind: cognition, affection, and conation
J. Hist. Behav. Sci.
Data Science
Xunzi, a Translation and Study of the Complete Works
Metaphors We Live by
3D data management: Controlling data volume, velocity and variety
Data as a resource: properties, implications, and prescriptions
MIT Sloan Manag. Rev.
An Introduction to Psychology
Three-Way Decision and Big Data Analysis
The wisdom web: new challenges for web intelligence (WI)
J. Intell. Inf. Syst.
Cited by (46)
A lattice-theoretic model of three-way conflict analysis
2024, Knowledge-Based SystemsThree-way clustering: Foundations, survey and challenges
2024, Applied Soft ComputingAnalysis of core attribute and approximate reduct based on the three-way decision
2024, Applied Soft ComputingThe Dao of three-way decision and three-world thinking
2023, International Journal of Approximate Reasoning