Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring the Impact of Personality Traits on Conversational Recommender Systems: A Simulation with Large Language Models

Xiaoyan Zhao, Yang Deng, Wenjie Wang, Hongzhan lin, Hong Cheng, Rui Zhang, See-Kiong Ng, Tat-Seng Chua

TL;DR

This paper tackles how user personality shapes outcomes in conversational recommender systems by introducing PerCRS, an LLM-based personality-aware user simulation driven by the Big Five framework for CRS (BF4CRS). It couples a personality-aligned user agent with a persuasion-capable system agent and evaluates them with a multi-aspect protocol across four data domains, revealing that LLMs can generate trait-consistent user responses and that personality traits meaningfully influence CRS effectiveness and strategy selection. Key findings show that traits such as Agreeableness and Extraversion drive higher success and engagement, while Neuroticism hampers persuasiveness, and that Emotional Resonance broadly outperforms other strategies. The work demonstrates the feasibility of simulating personality in CRSs to study interaction dynamics at scale, offering practical guidance for tailoring persuasion strategies and highlighting avenues for future safety-aware, multi-model personality research.

Abstract

Conversational Recommender Systems (CRSs) engage users in multi-turn interactions to deliver personalized recommendations. The emergence of large language models (LLMs) further enhances these systems by enabling more natural and dynamic user interactions. However, a key challenge remains in understanding how personality traits shape conversational recommendation outcomes. Psychological evidence highlights the influence of personality traits on user interaction behaviors. To address this, we introduce an LLM-based personality-aware user simulation for CRSs (PerCRS). The user agent induces customizable personality traits and preferences, while the system agent possesses the persuasion capability to simulate realistic interaction in CRSs. We incorporate multi-aspect evaluation to ensure robustness and conduct extensive analysis from both user and system perspectives. Experimental results demonstrate that state-of-the-art LLMs can effectively generate diverse user responses aligned with specified personality traits, thereby prompting CRSs to dynamically adjust their recommendation strategies. Our experimental analysis offers empirical insights into the impact of personality traits on the outcomes of conversational recommender systems.

Exploring the Impact of Personality Traits on Conversational Recommender Systems: A Simulation with Large Language Models

TL;DR

This paper tackles how user personality shapes outcomes in conversational recommender systems by introducing PerCRS, an LLM-based personality-aware user simulation driven by the Big Five framework for CRS (BF4CRS). It couples a personality-aligned user agent with a persuasion-capable system agent and evaluates them with a multi-aspect protocol across four data domains, revealing that LLMs can generate trait-consistent user responses and that personality traits meaningfully influence CRS effectiveness and strategy selection. Key findings show that traits such as Agreeableness and Extraversion drive higher success and engagement, while Neuroticism hampers persuasiveness, and that Emotional Resonance broadly outperforms other strategies. The work demonstrates the feasibility of simulating personality in CRSs to study interaction dynamics at scale, offering practical guidance for tailoring persuasion strategies and highlighting avenues for future safety-aware, multi-model personality research.

Abstract

Conversational Recommender Systems (CRSs) engage users in multi-turn interactions to deliver personalized recommendations. The emergence of large language models (LLMs) further enhances these systems by enabling more natural and dynamic user interactions. However, a key challenge remains in understanding how personality traits shape conversational recommendation outcomes. Psychological evidence highlights the influence of personality traits on user interaction behaviors. To address this, we introduce an LLM-based personality-aware user simulation for CRSs (PerCRS). The user agent induces customizable personality traits and preferences, while the system agent possesses the persuasion capability to simulate realistic interaction in CRSs. We incorporate multi-aspect evaluation to ensure robustness and conduct extensive analysis from both user and system perspectives. Experimental results demonstrate that state-of-the-art LLMs can effectively generate diverse user responses aligned with specified personality traits, thereby prompting CRSs to dynamically adjust their recommendation strategies. Our experimental analysis offers empirical insights into the impact of personality traits on the outcomes of conversational recommender systems.

Paper Structure

This paper contains 29 sections, 6 equations, 7 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Conversational Recommender System.
  4. Personality and LLMs.
  5. Methodology
  6. Personality Generator
  7. Personality-aware User Configuration
  8. CRS Simulation
  9. Experimental Setup
  10. Datasets
  11. Evaluation Metrics
  12. Evaluation of Personality Simulation Consistency.
  13. Evaluation of CRSs.
  14. Implementation Details
  15. Experimental Results
  16. ...and 14 more sections

Figures (7)

  • Figure 1: Different user shows different personality traits in CRS. The discussion process about the recommended item is omitted for brevity.
  • Figure 2: Overview of our PerCRS framework. This framework simulates personalized CRSs by personality-aware user configuration, where factuality source is to ensure the credibility of conversation contents. The complete conversation content is presented in Appendix \ref{['app:detail_figure_table']}.
  • Figure 3: Comparison of personality trait dimensions across five metrics (SR, GSR, PRS, SCR, TCR), highlighting the differences between with/without persuasion conditions. The raw results are provided in Appendix \ref{['app:detail_figure_table']}.
  • Figure 4: The frequency of persuasion strategies adopted by the system for user agents with different personality traits.
  • Figure 5: Prompt for the user agent with specified personality traits.
  • ...and 2 more figures