Understanding Long-Term Dynamics of Individual Metro Usage: A Hidden Semi-Markov State Framework with Survival Analysis

2026-06-17 · Source: stat.ML updates on arXiv.org · Field: Transportation & Mobility — Public Transportation & Urban Mobility, Mobility Services & Technology · Depth: Expert, extended

Summary

A new state-based lifecycle modeling framework integrates Hidden Semi-Markov Models (HSMM) with discrete-time survival analysis to characterize individual metro usage evolution over multi-year horizons. This approach addresses gaps in existing methods that treat mobility patterns as static or focus on short-term variability. Applied to four years of Shanghai metro smart card data (2021–2024) from 120,000 users, the framework identifies five robust mobility states and a directional transition hierarchy centered on an occasional-usage gateway state. The analysis reveals that exit hazard is state-dependent but duration-independent, while re-entry hazard decays sharply with inactivity length, offering a comprehensive view of transit participation lifecycles.

Key takeaway

For transit operators and planners aiming to improve ridership forecasting and passenger retention, you should adopt dynamic lifecycle monitoring over static user segmentation. This framework allows you to identify users vulnerable to disengagement based on their inferred behavioral states and understand how participation evolves, enabling the design of more targeted retention and reactivation strategies.

Key insights

The framework models individual metro usage lifecycles by integrating latent mobility states with exit and re-entry hazards.

Principles

• Individual mobility patterns are dynamic, not static.
• HSMMs explicitly model state dwell times.
• Disengagement and return risks are state-dependent.

Method

Sequentially integrate HSMM for latent state inference and trajectory decoding (Viterbi algorithm) with discrete-time survival models (logistic regression) for state-dependent exit/re-entry hazard estimation.

In practice

• Identify users vulnerable to disengagement.
• Design targeted retention strategies.
• Monitor dynamic user lifecycle evolution.

Topics

• Smart Card Data
• Mobility Lifecycle
• Hidden Semi-Markov Models
• Survival Analysis
• Transit Planning
• Passenger Retention

Best for: AI Scientist, Research Scientist, Data Scientist

Related on AIssential

• Fusing Urban Structure and Semantics: A Conditional Diffusion Model for Cross-City OD Matrix Generation — Fusing urban semantics and spatial constraints via a conditional diffusion model enhances cross-city OD matrix generation accuracy and generalizability.
• Psy-Chronicle:A Structured Pipeline for Synthesizing Long-Horizon Campus Psychological Counseling Dialogues — Long-horizon psychological counseling data can be synthesized by modeling student profiles, temporal stress events, and cross-session memory.
• Planning, Scheduling, and Behavior in EV Charging Systems: A Critical Survey and Trilemma Framework — EV charging system design faces a "PSB trilemma" balancing planning, scheduling, and user behavior fidelity with computational tractability.
• markov chains - explained #maths #statistics #mathematics #machinelearning #datascience — Markov chains model systems where future states depend only on the present state, defined by transition probabilities.
• Learning Survival Models with Right-Censored Reporting Delays — Jointly modeling event occurrence and report timing addresses right-censored reporting delays for timely risk evaluation.
• Learning Demographic-Conditioned Mobility Trajectories with Aggregate Supervision — ATLAS generates demographically realistic mobility trajectories using weak supervision from aggregate regional data.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by stat.ML updates on arXiv.org.