An OLS and GMM Combined Algorithm for Text Analysis for Heterogeneous Impact in Online Health Communities

Abstract

The increase of doctors' activity in online health communities (OHCs) plays a decisive role in their development. Although the literature on the determinants of doctors' online activities has received considerable attention, the impact of illness severity on these factors remains rare. A network externality analytical framework is constructed to explain the factors (that is, responsiveness, involvement, word-of-mouth, incentives, price, titles and gender) affecting online doctors' behavior, and assess whether factors differ by. By developing text analysis of 4916 doctors' data from a Chinese OHC, this paper applies ordinary least squares (OLS) and General Method of Moments (GMM) to analyze whether the determinants are equal across serious, moderate, and mild illnesses. Our experiment results find that the determinants affecting doctors' online service activity substantially differ across illness severity. Experiments prove the effectiveness of the proposed OLS and GMM methods and demonstrate that they are applicable in online medical field.

Keywords

• doctor activity
• GMM
• illness severity
• network externalities
• OLS
• online health communities