Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MISApp: Multi-Hop Intent-Aware Session Graph Learning for Next App Prediction

Yunchi Yang, Longlong Li, Jianliang Wu, Cunquan Qu

Abstract

Predicting the next mobile app a user will launch is essential for proactive mobile services. Yet accurate prediction remains challenging in real-world settings, where user intent can shift rapidly within short sessions and user-specific historical profiles are often sparse or unavailable, especially under cold-start conditions. Existing approaches mainly model app usage as sequential behavior or local session transitions, limiting their ability to capture higher-order structural dependencies and evolving session intent. To address this issue, we propose MISApp, a profile-free framework for next app prediction based on multi-hop session graph learning. MISApp constructs multi-hop session graphs to capture transition dependencies at different structural ranges, learns session representations through lightweight graph propagation, incorporates temporal and spatial context to characterize session conditions, and captures intent evolution from recent interactions. Experiments on two real-world app usage datasets show that MISApp consistently outperforms competitive baselines under both standard and cold-start settings, while maintaining a favorable balance between predictive accuracy and practical efficiency. Further analyses show that the learned hop-level attention weights align well with structural relevance, offering interpretable evidence for the effectiveness of the proposed multi-hop modeling strategy.

MISApp: Multi-Hop Intent-Aware Session Graph Learning for Next App Prediction

Abstract

Predicting the next mobile app a user will launch is essential for proactive mobile services. Yet accurate prediction remains challenging in real-world settings, where user intent can shift rapidly within short sessions and user-specific historical profiles are often sparse or unavailable, especially under cold-start conditions. Existing approaches mainly model app usage as sequential behavior or local session transitions, limiting their ability to capture higher-order structural dependencies and evolving session intent. To address this issue, we propose MISApp, a profile-free framework for next app prediction based on multi-hop session graph learning. MISApp constructs multi-hop session graphs to capture transition dependencies at different structural ranges, learns session representations through lightweight graph propagation, incorporates temporal and spatial context to characterize session conditions, and captures intent evolution from recent interactions. Experiments on two real-world app usage datasets show that MISApp consistently outperforms competitive baselines under both standard and cold-start settings, while maintaining a favorable balance between predictive accuracy and practical efficiency. Further analyses show that the learned hop-level attention weights align well with structural relevance, offering interpretable evidence for the effectiveness of the proposed multi-hop modeling strategy.
Paper Structure (24 sections, 37 equations, 4 figures, 10 tables)

This paper contains 24 sections, 37 equations, 4 figures, 10 tables.

Table of Contents

  1. Introduction
  2. RELATED WORK
  3. METHODOLOGY
  4. Framework Overview
  5. Multi-Hop Relation Graph Module
  6. Contextual Information Embedding
  7. Spatial Context Construction
  8. Spatial Context Embedding
  9. Cross-Modal Gated Fusion
  10. Cross-modal interaction
  11. Gated fusion
  12. Inter-Session Intent Transition Module
  13. EXPERIMENTS
  14. Dataset
  15. Experimental Settings
  16. ...and 9 more sections

Figures (4)

  • Figure 1: The overall framework of MISApp: the model integrates app, temporal and spatial context embedding through cross-modal gated fusion (CMGF) module, constructs multi-hop session graph, and obtains multi-hop structural features through lightgcn attention network. The generated graph embedding is combined with the immediate intent and processed by the transformer layer, which captures the sequential dependencies necessary for predicting the next app.
  • Figure 2: App prediction performance under varying immediate intent window sizes on the Tsinghua App Usage dataset. A. ACC@K evaluation results. B. MRR@K evaluation results.
  • Figure 4: Hop-level weights and structural correlation mechanism for Seq-1. Each app is annotated with both the app ID and its corresponding category label.
  • Figure 5: Comparison of Top-1 prediction results and target probabilities under original and perturbed structural settings. $\Delta$ denotes the drop in target prediction probability.