Application of 2-gram to Obtain Factor Scores of Statements Posted at Q&A Sites
Yuya Yokoyama, Teruhisa Hochin, Hiroki Nomiya · 2021
In order to solve the issues of mismatches between the intentions of questioners and respondents of Question and Answer (Q&A) sites, nine factors of impressions for Q&A statements have been obtained through impressive experiments. Factor scores have been then estimated through multiple regression analysis by using the feature values of statements, e.g. syntactic information, etc. The factor scores obtained have been subsequently used for inspecting the possibility of detecting respondents who are expected to post appropriate answer to a newly posted question. However, the method so far has been largely dependent on the syntactic information extracted through morphological analysis. In addition, the number of explanatory variables used for estimating factor scores has been considerably numerous and complicated. Therefore, in this paper, factor scores are estimated using the feature values of syntactic information are extracted through 2-gram, instead of morphological analysis. The feature values based on 2-gram and those other than the syntactic information are set as explanatory variable. As a result of another multiple regression analysis, estimation accuracy with 2-gram shows almost equal to or better than that with morphological analysis for all the nine factors. Moreover, monadic term is sufficient in estimating factor scores, leading to much smaller samples of explanatory variables.