Construction of Employer Reputation Evaluation Dimension of Crowdsourcing Platform based on Reviews Text Mining

Abstract:
Reputation evaluation mechanism is the basic guarantee for the healthy, orderly, and rapid development of crowdsourcing activities. Aiming at the lack of the employer reputation evaluation mechanism in crowdsourcing platforms, an employer reputation evaluation dimension based on text data mining is proposed. Taking the reviews text of the workers of a crowdsourcing platform as a case set, this paper uses the TF-IDF algorithm to extract the text features and analyzes some factors affecting the employer's reputation that the workers pays attention to and attaches importance to through keyword co-occurrence and social network analysis. Latent Dirichlet Allocation (LDA) is used for text topic clustering, and extract eight dimensions that effect the reputation of employers, namely explicit demand, smooth communication, timely payment, specialty, patient and meticulous, trust and consideration, cooperation experience, cooperation willingness. This study is of great significance to improve the employer reputation evaluation mechanism of the crowdsourcing platform, truly feedback on the employer reputation, and improve the performance of crowdsourcing tasks.