Research on An Electronic Map Retrieval Algorithm Based on Big Data

Jian Li, Rujing Yao, Shiji Zhu2017 · DOI

Abstract

Today, big data plays an important role. Electronic map has great research value as the core part of the application of big data. Electronic map retrieval is a hot spot in the research field of electronic map. This paper makes some improvement to the defect in traditional ZZL algorithm of electronic map retrieval algorithm that the same tail character string is initialized every time when we retrieve something so that the algorithm performance is reduced. We introduce clustering analysis in the traditional ZZL algorithm and propose an improved ZZL algorithm. The experimental results show that the improved ZZL algorithm has improved the retrieval efficiency by 37.1% on average compared with the traditional ZZL algorithm in the retrieval of geographic name in electronic map retrieval. The improved ZZL algorithm has an exploring significance to the development of the retrieval of electronic map.

Introduction

In the era of big data, electronic maps have evolved from simple visual aids into massive, searchable databases of geographical entities. Efficient string matching is the cornerstone of these search functionalities. However, traditional methods often struggle with the sheer volume of data in modern Geographic Information Systems (GIS).

The traditional ZZL algorithm, while specialized for string matching, suffers from a "redundant initialization" bottleneck. When searching for multiple entities sharing common suffixes (e.g., "Province," "City," or "Road"), the algorithm re-processes these shared "tail character strings" for every query, leading to significant computational waste and reduced performance.

Methodology: The Improved ZZL Algorithm

To address these inefficiencies, this research proposes an Improved ZZL Algorithm that incorporates clustering analysis and shifts the indexing focus from the start of a string to its tail.

Clustering Analysis

Instead of treating each search query in isolation, the improved algorithm groups pattern strings into clusters based on their generic terms. By categorizing entities that share common suffixes, the algorithm can effectively pre-sort the search space.

Tail-First Preprocessing

Unlike the traditional ZZL method that indexes the first character of a pattern, the improved version identifies the positions of "tail character strings" within the main database. These positions are stored in a dictionary, allowing the algorithm to bypass redundant initializations for strings with shared endings.

Optimized Matching Process

During retrieval, the algorithm utilizes a pointer system to navigate the pre-indexed dictionary. It then applies the standard Brute Force (BF) algorithm to verify matches only at these high-probability locations, significantly narrowing the search field and saving computational cycles.

Experimental Results

The algorithm was evaluated using geographical name datasets from Jilin, Hunan, and Hebei provinces across two scales: a small-scale set of 2,189 pattern strings and a large-scale set of 82,940 pattern strings.

Performance comparison between the Improved ZZL Algorithm and the Traditional ZZL Algorithm across different data scales (2,189 and 82,940 pattern strings).

As shown in the experimental data, the Improved ZZL algorithm consistently outperforms the traditional method in terms of processing time while maintaining high accuracy levels. For instance, in the large-scale test for Jilin Province, the retrieval time dropped from 88.6 seconds to 53.6 seconds. On average, the improved algorithm achieved a 37.1% increase in retrieval efficiency.

Accuracy and precision were maintained at over 90% across most tests. Even as the dataset size increased forty-fold, the accuracy remained robust, proving the algorithm's stability for big data applications.

Conclusion

By integrating clustering analysis and optimizing the handling of tail character strings, the Improved ZZL algorithm offers a robust solution for big data electronic map retrieval. The results demonstrate that the algorithm significantly reduces computational overhead without sacrificing reliability. Future work will focus on further optimizing preprocessing for high-density clusters and exploring applications in other big data retrieval domains.

Summary updated

Abstract

Today, big data plays an important role. Electronic map has great research value as the core part of the application of big data. Electronic map retrieval is a hot spot in the research field of electronic map. This paper makes some improvement to the defect in traditional ZZL algorithm of electronic map retrieval algorithm that the same tail character string is initialized every time when we retrieve something so that the algorithm performance is reduced. We introduce clustering analysis in the traditional ZZL algorithm and propose an improved ZZL algorithm. The experimental results show that the improved ZZL algorithm has improved the retrieval efficiency by 37.1% on average compared with the traditional ZZL algorithm in the retrieval of geographic name in electronic map retrieval. The improved ZZL algorithm has an exploring significance to the development of the retrieval of electronic map.
BibTeX
@misc{li2017research,
  author = {Jian Li and Rujing Yao and Shiji Zhu},
  title = {{Research on An Electronic Map Retrieval Algorithm Based on Big Data}},
  year = {2017},
  doi = {10.22323/1.299.0055},
}
Figure