Work

Intelligent Classification System for Rail Transit Engineering Quotas Based on K-Means Clustering and Random Forest

Year: 2026

Type: conference-paper

Abstract: Efficient and accurate quota classification is fundamental to rapid cost estimation in rail transit projects. To address the limitations of manual, experience-based quota matching, this study proposes an intelligent classification system that integrates unsupervised and supervised learning. First, an 11-dimensional feature vector characterizing labor intensity, cost structure, and geological attributes was constructed from 128 real-work quota records. K-Means clustering was then applied to discover inherent data structures, automatically grouping quotas into five distinct categories such as "Material-Dominant" and "Machinery-Intensive." These cluster labels subsequently trained a Random Forest classifier. The integrated system achieved a classification accuracy of 92.3% on a held-out test set. Feature importance analysis indicated that geological condition, labor density, and unit type were the most discriminative factors. This clustering-classification framework provides a promising data-driven pathway for automating quota identification and supporting rapid cost estimation in rail transit engineering.

Source: International Conference on Innovation Management and Information Systems (ICIMIS)

DOI: 10.1145/3821966.3822059

Authors: Shuai He, Jie Li, Zhiyuan Li, Yanxu Guo

Institutions: Shandong University, Weifang University

Language: English

Topic: Data Mining and Machine Learning Applications

Subfield: Information Systems

Field: Computer Science

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Sustainable cities and communities

Open Access status: hybrid

OpenAlex ID: W7204990795

FWCI:

Cites: 13

Cited by:

Related to: 10