Research on the Intelligent Distinction Mechanism of Tourism Preference and Purpose Based on the DIKWP × DIKWP Model
Yucong Duan
International Standardization Committee of Networked DIKWPfor Artificial Intelligence Evaluation(DIKWP-SC)
World Artificial Consciousness CIC(WAC)
World Conference on Artificial Consciousness(WCAC)
(Email: duanyucong@hotmail.com)
Abstract
This paper, based on the DIKWP model (Data-Information-Knowledge-Wisdom-Purpose) proposed by Yucong Duan, explores the distinction mechanism and importance of tourism preference versus tourism intention in intelligent tourism systems. The DIKWP model introduces a "Purpose" layer to the traditional DIKW cognitive hierarchy and adopts a network structure to achieve bidirectional feedback among layers. Through this model, we conduct an in-depth analysis of the different logical structures and generation mechanisms of tourist preferences and intentions. Preferences are shown to be accumulated tendencies at the levels of data, information, knowledge, and wisdom, which can be mapped to specific layers or simple paths within the DIKWP model (e.g., the transformation from data to knowledge). In contrast, intentions are manifested through more complex multi-layer connections and reasoning closed-loops, reflecting an active cognitive process driven by "Purpose (P)". This paper focuses on the mechanism of migrating from an abstract conceptual space to a semantic space to avoid biases caused by rigid definitions of "concepts." Leveraging the DIKWP×DIKWP interaction network, we transform the concept of preference in tourism contexts into a semantic representation, dynamically aligning the understanding of meaning between tourists and intelligent systems. Based on theoretical arguments and model construction, this paper proposes an architectural vision for an intelligent tourism management system, including DIKWP modules for profiling tourists and tourism products, an intelligent recognition algorithm for preference-intention mapping, and personalized recommendation and interaction mechanisms. Finally, through case studies and future research prospects, we discuss the feasibility and potential benefits of applying the DIKWP×DIKWP model to the distinction between tourism preference and intention. The research indicates that introducing the DIKWP cognitive model can enrich the semantic layers of personalized tourism services, effectively avoid semantic misunderstandings, and enhance the system's ability to understand and explain tourists' true needs, providing a new theoretical perspective and technical path for the construction of intelligent tourism systems.
Keywords: Tourism Preference; Tourism Intention; DIKWP Model; Conceptual Space; Semantic Space; Personalized Intelligent Tourism
1. Introduction
In the fields of smart tourism and personalized recommendation, accurately grasping tourists' preferences and intentions is one of the key challenges in building high-quality tourism management systems. Preference usually refers to a tourist's relatively stable inclination or liking, such as a love for natural scenery or urban culture. Intention, on the other hand, refers to a tourist's goal or plan in a specific context, such as planning a vacation to a certain island next month. The two are closely related but fundamentally different: preference focuses on long-term accumulated interests, while intention emphasizes current motivation and the drive for decision-making action. However, in many existing tourism systems and research, preference and intention are often used interchangeably and vaguely, lacking a clear distinction. For example, in practical industry applications, "travel intention" is sometimes equated with "destination preference." This can lead to recommendation systems that only push content based on a user's past preferences, failing to accurately capture their current, true travel purpose.
Traditional tourism recommendations mainly rely on big data and statistical learning to mine preferences from users' historical behavior and then predict their possible travel destinations or product choices. This type of surface-level customization can generally adjust content presentation according to user preferences, such as recommending attractions or hotels that match their interests. However, this method has its limitations: it focuses on behavioral patterns and symbolic matching, with less consideration for the issue of semantic interpretation consistency. In other words, the system can guess "what the user might like," but it is difficult to ensure that it "understands why the user likes it and what they want at this moment." When different tourists have different understandings of the same concept, preference matching alone may lead to communication errors. For example, "a thrilling adventure" has different meanings in the minds of different tourists. A tourist who prefers cultural exploration and one who prefers extreme sports have significantly different understandings of the concept of "adventure." If the system simply recommends high-intensity outdoor activities based on the keyword "adventure," it may not match the true intention of the former, thus reducing satisfaction.
Therefore, deeply distinguishing between preference and intention and achieving alignment at the semantic level is crucial for the precise decision-making of intelligent tourism systems. This is the starting point of this research. We introduce the DIKWP cognitive model as a theoretical basis to attempt to solve the problem of preference-intention distinction with a systematic framework. The DIKWP model is composed of five layers—Data, Information, Knowledge, Wisdom, and Purpose. By adding the top-level factor of "Purpose," it constructs a goal-driven cognitive structure. Compared to traditional linear information processing models, the DIKWP model emphasizes networked interaction and closed-loop feedback: higher-level wisdom and purpose can influence the interpretation of lower-level data and information, and new lower-level information can in turn modify higher-level cognition, achieving continuous iterative optimization. This feature provides us with a unified language to describe the cognitive processes of both humans and AI, allowing us to study the issues of tourism preference and intention in a deeper semantic space.
Based on the DIKWP theory proposed by Professor Yucong Duan, and combined with the tourism context, this paper aims to:
1.Clarify the cognitive and logical differences between tourism preference and intention, explaining how preference originates from information processing at different levels, and how intention acts as a driving factor to guide the decision-making process.
2.Propose a preference-intention distinction mechanism based on the DIKWP×DIKWP model, which avoids the biases brought by rigid definitions of abstract concepts through the transformation from conceptual space to semantic space, achieving personalization in understanding tourist semantics.
3.Construct an architectural framework for an intelligent tourism management system, exploring possible solutions for applying the DIKWP model to tourist profiling and tourism product profiling, and designing preference/intention-driven recommendation and interaction mechanisms.
4.Preliminarily verify the effectiveness of the above model through case studies or simulation experiments, and look forward to future empirical research plans.
Through this work, we hope to provide a new theoretical perspective and technical path for the field of tourism management, enhancing the depth of understanding and the effectiveness of interaction of artificial intelligence regarding tourist needs.
2. Theoretical Basis: The DIKWP Model and Tourism Preference and Intention
2.1 Overview of the DIKWP Model
The DIKWP model is a five-layer cognitive model of "Data-Information-Knowledge-Wisdom-Purpose," proposed by Yucong Duan and others in the study of artificial intelligence semantics. The model originates from the classic DIKW system, but the key difference lies in the introduction of "Purpose" as a new top-level factor and a reinterpretation of the hierarchical relationships. The traditional DIKW pyramid describes the cognitive processing process in a linear, unidirectional way, with data being abstracted step-by-step into information, knowledge, and finally wisdom. However, Duan et al. pointed out that DIKW lacks consideration for purpose/intention, which limits its ability to characterize the proactiveness of intelligent agents. To address this, DIKWP adds a Purpose layer above wisdom and endows the model with a networked structure.
In the DIKWP model, the meanings of each layer are as follows:
·Data (D): Objective, raw facts and perceptual inputs, such as statistical numbers, sensor readings, text records, etc. The Data layer corresponds to specific observations and discrete materials, forming the basis of cognition.
·Information (I): Data that has been organized and endowed with a certain context or grammatical structure. For example, processing raw data into descriptive statements, tables, images, etc., to make it meaningful to the cognitive subject. The Information layer represents the initial interpretation and connection of data.
·Knowledge (K): Patterns, relationships, and judgments summarized from information, which can be regarded as a higher level of understanding. The Knowledge layer includes models, rules, conceptual systems (such as ontologies), etc., which generalize information and give it causal or logical structure.
·Wisdom (W): Insights or decision-making strategies formed on the basis of rich knowledge, combined with experience and values. The Wisdom layer embodies judgment in complex situations and the principles for trade-offs, often involving fuzzy and comprehensive cognition (such as rules of thumb, professional intuition).
·Purpose (P): The motivation, goal, or driving force of the cognitive subject. As the highest layer, the Purpose layer determines "why" the cognitive process occurs and what goal it is directed towards. It represents the desired state that the intelligent agent is trying to achieve or the problem it is trying to solve.
It is worth noting that the DIKWP model does not simply connect these five elements in series but integrates them into an organic whole through networked interaction. There are not only bottom-up abstraction and refinement relationships between the layers but also top-down guidance and feedback relationships. This dynamic, white-box transparent semantic space allows for the clear representation and explanation of content at each level and the tracking of their interaction paths, providing a foundation for personalized, semantic intelligent applications.
2.2 Tourism Preference and Intention: Conceptual Connotation and Distinction
Tourism preference typically refers to a tourist's lasting fondness for a certain type of travel experience or element. Preferences often stem from an individual's past experiences, personality traits, and values, and they have relative stability and are long-term.
Tourism intention refers to the travel motivation and plan formed by a tourist at a specific point in time or in a specific context, manifested as the willingness to take a certain travel action. Intention usually contains a clear target object or result.
To distinguish between the two more intuitively, Table 1 lists a comparison of tourism preference and tourism intention in several aspects:
Aspect |
Tourism Preference |
Tourism Intention |
Directionality |
Points to categories or attributes (e.g., liking nature, culture, food). |
Points to specific actions or goals (e.g., going to a place, participating in an activity). |
Time Dimension |
Usually long-term and stable. |
Has immediacy and changes with the situation. |
Formation Process |
Shaped by accumulated knowledge and experience. |
Generated in real-time under specific information stimuli and cognitive reasoning. |
Role in Decision-Making |
Acts as a filter and weighting factor. |
The direct driving force of a decision. |
Example |
A long-term preference for historical and cultural tourism. |
An intention to "visit the Terracotta Army exhibition in Xi'an next month." |
Distinguishing between preference and intention is crucial for building intelligent tourism management systems because it allows for a more accurate understanding of user needs, informs different levels of recommendation strategies, improves the interactive experience, and deepens the theoretical understanding of tourist behavior.
3. DIKWP Logical Structure Analysis of Tourism Preference and Intention
3.1 Mapping of Preference in the DIKWP Model
Tourist preferences can be seen as the reinforced expression of certain elements or paths in the DIKWP semantic space. Specifically, the formation and expression of preferences are related to:
·Accumulated tendencies at the Knowledge/Wisdom layers: Preferences often exist as knowledge in an individual's mind.
·Attention selection at the Data/Information layers: Tourists with different preferences show selective attention when perceiving and acquiring information.
·Position in the conceptual space: Preferences sometimes appear as conceptual labels, like "backpacker" or "artsy youth."
·Path from Data to Knowledge: The formation of preferences often involves repeated transformation and reinforcement along the "Data → Information → Knowledge" path.
In summary, tourism preference can be mapped to a relatively static semantic structure in the DIKWP model.
3.2 Mapping of Intention in the DIKWP Model
Compared to the static nature of preference, intention in the DIKWP model is a dynamic, pervasive element. As the driving factor at the highest level, intention not only represents a cognitive goal itself but also guides the entire DIKWP process. We can understand the embodiment of intention in the model from the following aspects:
·Explicit representation at the Purpose layer: Intention has a direct correspondence in the DIKWP model—the Purpose (P) layer.
·Connection and constraint of multi-layer paths: One of the roles of intention is to constrain and organize cognitive paths.
·Cognitive closed-loop and reasoning: Due to the existence of intention, DIKWP processing often exhibits the characteristics of reasoning and problem-solving.
·Context dependence and role differences: Intention is contextual and can be completely different in different scenarios, times, and roles, even for the same person.
·Semantic alignment and misunderstanding detection: In human-computer interaction, identifying the user's true intention is a key step for the system to understand the input.
In summary, intention in the DIKWP model is both a high-level entity (a P-layer node) and a logical thread running through all layers. This clear distinction lays the foundation for designing a preference-intention distinction mechanism.
4. Preference-Intention Distinction Mechanism Based on the DIKWP×DIKWP Model
This section explores how two cognitive subjects (or two cognitive spaces) can achieve semantic alignment and transformation through DIKWP×DIKWP interaction, thereby achieving the goal of distinguishing between preference and intention.
4.1 DIKWP×DIKWP Interaction Network Architecture
The DIKWP×DIKWP network, proposed by Duan et al. in recent work, is a dual-cycle interactive architecture containing two sets of DIKWP semantic networks. Intuitively, we can see it as two intelligent agents (or two semantic spaces), each with its own five layers of semantic nodes and internal connections, with a series of cross-subject mapping connections established between them. The architecture specifies how each layer should interface to ensure accurate transmission. This interactive network essentially represents the external manifestation of both parties building a "common ground."
4.2 Migration from Conceptual Space to Semantic Space
As mentioned earlier, preferences are often mentioned at the conceptual level, and intentions are also often expressed with linguistic concepts (e.g., "I want an adventure"). However, concepts may correspond to different associations and importance in different semantic spaces. Therefore, we must avoid rigid definitions of concepts in system design and instead understand the user's true meaning through concept-to-semantic migration. This is precisely the problem that the DIKWP model, combined with the methods of semantic mathematics, aims to solve. This migration involves two steps: concept interpretation enhancement and semantic alignment transmission.
4.3 Intelligent Recognition Process for Preference and Intention
With the above framework, we can design a process for distinguishing between preference and intention within an intelligent tourism system:
1.User Profile Initialization (Offline Phase): The system builds a five-layer DIKWP profile for the user.
2.Scene Detection and Context Activation: The system determines whether the user is in a browsing state (no clear intention) or a goal-oriented state.
3.Preference Analysis: The system refers to the preference part of the user profile to filter or adjust the output content.
4.Intention Recognition: The system runs an intention recognition algorithm to parse the user's current expressed intention.
5.Interactive Clarification: If there is uncertainty in intention recognition, the system initiates a clarifying dialogue.
6.DIKWP×DIKWP Alignment Implementation: The system's knowledge base is matched against the user's model to find a solution that satisfies the intention.
7.Result Interpretation and Presentation: The system provides an explanation with the recommendation results, indicating how it has taken both the user's preferences and intentions into account.
Through this process, the system achieves a layered interpretation and separate treatment of preference and intention.
5. Intelligent Tourism Management System Architecture Design
Based on the above theories and mechanisms, we propose a conceptual plan for an intelligent tourism management system architecture that applies the DIKWP×DIKWP model to the design of various system components.
·Tourist DIKWP Semantic Profile: The tourist profile is upgraded to a five-layer semantic profile, characterizing the user's complete cognitive features from data perception to motivational intent.
·Tourism Product DIKWP Profile: Similarly, tourism products can be represented using the DIKWP model to characterize their features and scope of application.
·Preference-Intention Driven Recommendation and Interaction Mechanism: The core recommendation engine and interactive interface are designed around the dynamic combination of preference and intention, utilizing a dual-track strategy (preference track and intention track) and prioritizing semantic alignment in the human-computer interface.
6. Case Study and Discussion
To test and illustrate the effectiveness of the above model and architecture, we use a simulated travel decision scenario to demonstrate the role of the preference-intention distinction mechanism in a practical application. We consider two tourists with very different styles, Mr. A (an adventure enthusiast) and Ms. B (a family-oriented traveler), both interested in "Zhangjiajie" but with different intentions.
The system, equipped with their DIKWP profiles, is able to:
1.Understand their initial inquiries differently: For Mr. A, "Is it fun?" is interpreted as "Are there thrilling activities?" For Ms. B, it's interpreted as "Is it suitable for a family trip?"
2.Recommend different itineraries: Adventure tours for Mr. A and family-friendly tours for Ms. B.
3.Handle follow-up questions effectively: The system addresses Mr. A's questions about price differences (Data layer) and Ms. B's concerns about the physical demands of the trip (Wisdom layer).
4.Dynamically adjust to new information: When Mr. A mentions a friend with no experience wants to join, the system re-evaluates the chosen plan and suggests alternatives, recognizing the change in intention and constraints.
The case demonstrates the system's advantages in utilizing preference, recognizing intention dynamically, resolving conceptual ambiguity, providing explainability, and achieving a synergy between preference and intention. While challenges remain in real-world application, the deep semantic-driven approach shows great potential.
7. Conclusion
This research, based on Yucong Duan's DIKWP cognitive model, has conducted an in-depth exploration and framework design for the key issue in tourism management systems: how to intelligently distinguish and handle tourist preferences and travel intentions. We have clarified the conceptual differences, proposed a DIKWP mapping framework for both, and outlined an intelligent tourism system architecture based on the DIKWP×DIKWP interaction model.
The main theoretical contribution lies in providing a formal distinction between preference and intention within a cognitive framework, offering a novel perspective for understanding tourist decision-making. The practical implications are demonstrated through a simulated case, showing how this approach can lead to more accurate, personalized, and trustworthy recommendations, ultimately enhancing the user experience.
Future work will involve user studies to validate the model, development of a prototype system for empirical testing, optimization of the DIKWP×DIKWP semantic matching algorithms, and extending the application of this framework to other service domains. In conclusion, the preference-intention distinction mechanism based on the DIKWP model injects new vitality and intelligence into smart tourism management, paving the way for a new era of personalized services that are more aligned with human cognition and values.
A list of references was provided in the original document.
