TY - GEN
T1 - Using 4-gram to Obtain Factor Scores of Japanese Statements Posted at Q&A Sites
AU - Yokoyama, Yuya
AU - Hochin, Teruhisa
AU - Nomiya, Hiroki
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - With the aim of clearing 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 impression evaluation experiments. Then by means of multiple regression analysis, factor scores have been estimated by using the feature values of statements e.g., syntactic information, etc. The factor scores estimated and obtained have been subsequently employed for inspecting their potential to detect respondents who are likely to post appropriate answers to a posted question. However, our method has considerably depended on the syntactic information extracted through morphological analysis. Furthermore, this method has a critical downside of numerous explanatory variables and complicated multiple regression equation to estimate factor scores. Thus, we have shifted our principle to use N-gram on top of morphological analysis. In this paper, 4-gram is applied to the feature values, in a similar way as the previous analysis using 2-gram and 3-gram. Further analysis has shown that 4-gram is also applicable to our method. As a result of the analysis, 2-gram, 3-gram and 4-gram also outscore morphological analysis in terms of estimation accuracy; above all 2-gram shows the best accuracy. By and large, it could be suggested that N-gram as well as morphological analysis should be more sufficient than mere morphological analysis as the syntactic information of the feature values to estimate factor scores.
AB - With the aim of clearing 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 impression evaluation experiments. Then by means of multiple regression analysis, factor scores have been estimated by using the feature values of statements e.g., syntactic information, etc. The factor scores estimated and obtained have been subsequently employed for inspecting their potential to detect respondents who are likely to post appropriate answers to a posted question. However, our method has considerably depended on the syntactic information extracted through morphological analysis. Furthermore, this method has a critical downside of numerous explanatory variables and complicated multiple regression equation to estimate factor scores. Thus, we have shifted our principle to use N-gram on top of morphological analysis. In this paper, 4-gram is applied to the feature values, in a similar way as the previous analysis using 2-gram and 3-gram. Further analysis has shown that 4-gram is also applicable to our method. As a result of the analysis, 2-gram, 3-gram and 4-gram also outscore morphological analysis in terms of estimation accuracy; above all 2-gram shows the best accuracy. By and large, it could be suggested that N-gram as well as morphological analysis should be more sufficient than mere morphological analysis as the syntactic information of the feature values to estimate factor scores.
KW - Factor Score
KW - Multiple Regression Analysis
KW - N-gram
KW - Q&A Site
UR - https://www.scopus.com/pages/publications/85160934578
UR - https://www.scopus.com/pages/publications/85160934578#tab=citedBy
U2 - 10.1109/IIAI-AAI-Winter58034.2022.00016
DO - 10.1109/IIAI-AAI-Winter58034.2022.00016
M3 - Conference contribution
AN - SCOPUS:85160934578
T3 - Proceedings - 2022 13th International Congress on Advanced Applied Informatics Winter, IIAI-AAI-Winter 2022
SP - 25
EP - 31
BT - Proceedings - 2022 13th International Congress on Advanced Applied Informatics Winter, IIAI-AAI-Winter 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 13th IIAI International Congress on Advanced Applied Informatics Winter, IIAI-AAI-Winter 2022
Y2 - 12 December 2022 through 14 December 2022
ER -