Research on identification method for spatial master data of railroad system

Maolin Zhang, Tao Shen, Liang Huo, Yaodong Yang, Shiqi ZhouwenResearch Square 2024 · DOI

Abstract

Railway spatial master data is crucial asset data within the railway system, accurately reflecting the geographic locations and spatial information of core entities within the system, enabling railway companies to make a series of spatial decisions. In view of this, this article proposes a railway spatial master data recognition method based on the rough set-cloud model to identify spatial master data from massive railway data. This method first selects indicators for identification based on the characteristics of spatial data and master data. Then, it determines the weights of the identification indicators based on rough set theory. Next, it uses a cloud model to transform the indicators from qualitative to quantitative, calculate the membership degrees of the indicators, and generate an indicator cloud reflecting the characteristics of the indicators. Finally, by combining membership degrees with indicator weights and prioritizing maximum membership degrees, accurate identification of massive railway data is carried out. This method, compared to traditional recognition approaches, eliminates the subjective bias that expert scoring may introduce. It also increases the number of recognition indicators, enhancing accuracy and specificity in identification. The results indicate that the calculation method is consistent with the expected results, demonstrating the effectiveness and reliability of the method in the field of railway spatial master data identification.

Introduction

In the context of China’s expansive railway infrastructure—exceeding 155,000 kilometers—the volume of generated data is staggering. Amidst this vast "data swamp," identifying Spatial Master Data (SMD) is critical. SMD serves as the "golden record" for core entities such as stations, tracks, and bridges, which are vital for cross-departmental spatial decision-making and route planning.

Traditional identification methods, such as the Analytic Hierarchy Process (AHP), often fall short because they rely heavily on expert scoring. This introduces subjective bias, where results can vary significantly depending on the individual expert's perspective. To address this, researchers have developed a more objective, data-driven identification method based on a hybrid of Rough Set Theory and Cloud Models.

Methodology: A Hybrid Mathematical Framework

The proposed method utilizes a multi-stage pipeline to transform raw railway data into identified spatial master records. The process is designed to handle the inherent "fuzziness" of data definitions while maintaining mathematical objectivity.

1. Indicator Selection and Weighting

The researchers identified 10 key indicators to evaluate data, including uniqueness, shareability, and frequency of use. To avoid the "subjectivity trap" of expert-led weighting, Rough Set Theory was employed. This approach calculates weights based on the actual dependency and importance of attributes within the dataset itself, allowing the data to dictate which indicators are most significant.

2. Quantification via Cloud Models

Qualitative concepts like "high frequency of use" are often vague. The Cloud Model bridges the gap between human language and machine-readable mathematics by transforming these concepts into quantitative "cloud droplets." This is achieved using three parameters: Expectation (Ex), Entropy (En), and Hyper-entropy (He), which account for uncertainty and fuzziness.

3. Integrated Evaluation

A Forward Cloud Generator algorithm produces membership degrees for each indicator. These degrees are then combined with the weights derived from Rough Set Theory to produce a final classification (Levels I through V), where Level I represents the highest priority spatial master data.

Methodology flowchart for spatial master data identification

Experimental Validation

The model was validated using real-world data provided by a Chinese railroad company, supported by the China Railway Research Institute Group. The evaluation included core infrastructure (stations, switches, lines), engineering samples (tunnels, bridges), and non-spatial control entities (personnel records).

Key Results

  • Successful Identification: Rail Stations, Switches, Lines, Tunnels, and Bridges were all classified as Level I or II, confirming their status as critical Spatial Master Data.
  • Effective Filtering: Personnel records were correctly classified as Level V, demonstrating the model's ability to distinguish non-spatial data.
  • Precision: For the "Rail Station" sample, the model calculated a Level I membership of 0.3365, providing a clear mathematical basis for its classification.

Discussion and Future Outlook

By removing the "human in the loop" for weight calculation, this method ensures that the identification process is reproducible and objective. It treats uncertainty as a measurable variable rather than an error, which is particularly useful for complex infrastructure datasets.

However, some challenges remain. The process of discretizing data into "high, medium, and low" values requires further study to understand its sensitivity. Additionally, while more objective, the hybrid model is computationally intensive. Future research will focus on scaling this framework for real-time processing of millions of entities and adapting it to dynamic spatial data, such as real-time train locations.

Summary updated

Abstract

<title>Abstract</title> Railway spatial master data is crucial asset data within the railway system, accurately reflecting the geographic locations and spatial information of core entities within the system, enabling railway companies to make a series of spatial decisions. In view of this, this article proposes a railway spatial master data recognition method based on the rough set-cloud model to identify spatial master data from massive railway data. This method first selects indicators for identification based on the characteristics of spatial data and master data. Then, it determines the weights of the identification indicators based on rough set theory. Next, it uses a cloud model to transform the indicators from qualitative to quantitative, calculate the membership degrees of the indicators, and generate an indicator cloud reflecting the characteristics of the indicators. Finally, by combining membership degrees with indicator weights and prioritizing maximum membership degrees, accurate identification of massive railway data is carried out. This method, compared to traditional recognition approaches, eliminates the subjective bias that expert scoring may introduce. It also increases the number of recognition indicators, enhancing accuracy and specificity in identification. The results indicate that the calculation method is consistent with the expected results, demonstrating the effectiveness and reliability of the method in the field of railway spatial master data identification.
BibTeX
@inproceedings{zhang2024research,
  author = {Maolin Zhang and Tao Shen and Liang Huo and Yaodong Yang and Shiqi Zhouwen},
  title = {{Research on identification method for spatial master data of railroad system}},
  booktitle = {Research Square},
  year = {2024},
  doi = {10.21203/rs.3.rs-4453233/v1},
}
Figure