Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TFPS: A Temporal Filtration-enhanced Positive Sample Set Construction Method for Implicit Collaborative Filtering

Jiayi Wu, Zhengyu Wu, Xunkai Li, Rong-Hua Li, Guoren Wang

TL;DR

A novel temporal filtration-enhanced approach to construct a high-quality positive sample set and it is shown that TFPS can be integrated with various implicit CF recommenders or negative sampling methods to enhance its performance.

Abstract

The negative sampling strategy can effectively train collaborative filtering (CF) recommendation models based on implicit feedback by constructing positive and negative samples. However, existing methods primarily optimize the negative sampling process while neglecting the exploration of positive samples. Some denoising recommendation methods can be applied to denoise positive samples within negative sampling strategies, but they ignore temporal information. Existing work integrates sequential information during model aggregation but neglects time interval information, hindering accurate capture of users' current preferences. To address this problem, from a data perspective, we propose a novel temporal filtration-enhanced approach to construct a high-quality positive sample set. First, we design a time decay model based on interaction time intervals, transforming the original graph into a weighted user-item bipartite graph. Then, based on predefined filtering operations, the weighted user-item bipartite graph is layered. Finally, we design a layer-enhancement strategy to construct a high-quality positive sample set for the layered subgraphs. We provide theoretical insights into why TFPS can improve Recall@k and NDCG@k, and extensive experiments on three real-world datasets demonstrate the effectiveness of the proposed method. Additionally, TFPS can be integrated with various implicit CF recommenders or negative sampling methods to enhance its performance.

TFPS: A Temporal Filtration-enhanced Positive Sample Set Construction Method for Implicit Collaborative Filtering

TL;DR

A novel temporal filtration-enhanced approach to construct a high-quality positive sample set and it is shown that TFPS can be integrated with various implicit CF recommenders or negative sampling methods to enhance its performance.

Abstract

The negative sampling strategy can effectively train collaborative filtering (CF) recommendation models based on implicit feedback by constructing positive and negative samples. However, existing methods primarily optimize the negative sampling process while neglecting the exploration of positive samples. Some denoising recommendation methods can be applied to denoise positive samples within negative sampling strategies, but they ignore temporal information. Existing work integrates sequential information during model aggregation but neglects time interval information, hindering accurate capture of users' current preferences. To address this problem, from a data perspective, we propose a novel temporal filtration-enhanced approach to construct a high-quality positive sample set. First, we design a time decay model based on interaction time intervals, transforming the original graph into a weighted user-item bipartite graph. Then, based on predefined filtering operations, the weighted user-item bipartite graph is layered. Finally, we design a layer-enhancement strategy to construct a high-quality positive sample set for the layered subgraphs. We provide theoretical insights into why TFPS can improve Recall@k and NDCG@k, and extensive experiments on three real-world datasets demonstrate the effectiveness of the proposed method. Additionally, TFPS can be integrated with various implicit CF recommenders or negative sampling methods to enhance its performance.
Paper Structure (22 sections, 3 theorems, 11 equations, 7 figures, 4 tables, 1 algorithm)

This paper contains 22 sections, 3 theorems, 11 equations, 7 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Backgrounds
  4. Existing Methods and Their Defects
  5. Methodology
  6. Construct the Weighted Bipartite Graph
  7. Construct the Positive Sample Set
  8. Layered Subgraph Construction
  9. Layer Enhancement
  10. Model Optimization
  11. Theoretical Analysis
  12. Time Complexity
  13. Experiments
  14. Experimental Setup
  15. Datasets
  16. ...and 7 more sections

Key Result

theorem 1

Consider an implicit CF model trained with the BPR loss and a first-order optimizer (e.g., Adam). At each step, a negative item $p^- \sim f(\cdot\mid u,p)$ is sampled using the current parameters before the subsequent update. Let $s_{up}(\theta)$ denote the model score of user $u$ on item $p$ under Moreover, TFPS performs data-level reweighting by duplicating high-weight (recent) interactions, th

Figures (7)

  • Figure 1: Impact of data partitioning on model performance.
  • Figure 2: An overview of TFPS, where the darker the color of the edges, the higher the edge weight.
  • Figure 3: Recall@20 vs. wall-clock time (in seconds).
  • Figure 4: The impact of $\lambda$ and $n$ on Recall@30 and NDCG@30.
  • Figure 5: Impact of Positive Sample Distribution.
  • ...and 2 more figures

Theorems & Definitions (4)

  • theorem 1: Local one-step margin improvement and TFPS-induced update amplification
  • proof : Proof sketch
  • corollary 1: TFPS insight for test Recall@k via margin amplification
  • corollary 2: TFPS insight for test NDCG@k via a margin-monotone lower bound