CrowdPlanner: A Crowd-Based Route Recommendation System

TL;DR

CrowdPlanner is a new system that leverages human workers to evaluate and select optimal routes from multiple recommendations, improving route recommendation accuracy through optimized task design and worker selection.

Contribution

It introduces a novel crowd-based approach with optimized question generation and worker selection algorithms for route recommendation.

Findings

Effective task generation improves worker response quality.

Optimized worker selection enhances route recommendation accuracy.

System outperforms baseline methods in route quality assessment.

Abstract

CrowdPlanner -- a novel crowd-based route recommendation system has been developed, which requests human workers to evaluate candidates routes recommended by different sources and methods, and determine the best route based on the feedbacks of these workers. Our system addresses two critical issues in its core components: a) task generation component generates a series of informative and concise questions with optimized ordering for a given candidate route set so that workers feel comfortable and easy to answer; and b) worker selection component utilizes a set of selection criteria and an efficient algorithm to find the most eligible workers to answer the questions with high accuracy.