Frontiers in Educational Research, 2020, 3(7); doi: 10.25236/FER.2020.030720.
Chenhao Ma1, Shuxian Song2, Shun Xu3, Shangrui Xiao4
1 College of Information Science and Technology, Dalian Maritime University, Dalian Liaoning 116000, China
2 Maritime Institute, Dalian Maritime University, Dalian Liaoning 116000, China
3 College of Naval Architecture and Ocean Engineering, Dalian Maritime University, Dalian Liaoning 116000, China
4 College of Safety and Ocean Engineering, China University of Petroleum (Beijing), Beijing 100050, China
In order to draw up Sunshine company' s online sales strategy, enhance customers' satisfaction to the products, and predict Sunshine's future sales trends, we do the following work:First, during the performance of data preprocessing, we consider the data labels as variables, process abnormal data, and select meaningful variables based on data redundancy and association principles. Then we regularize the text review, calculate the frequency of emotional words, words-modifiers and negation words and analyze the emotional level of the review body qualitatively. By developing the correlation analysis model, and using the Spearman coefficient to measure the correlation between different variables, we find that there is a strong correlation between helpful votes and the total votes.Secondly, we group and rank emotional words and words modifiers, assign negation values to -1, consider the length of the text-review and the number of times that negation words appear to build the text-review emotion quantification model to calculate the emotional values of each review. And normalizing emotional values which range [0, 5]. After that, with the linear weighted sum method between the star ratings and the emotional values, we obtain a new score called 〖score(i)〗_new, and refer to the literature for a set of reasonable weight values.
Text- review emotion quantification model, Correlation analysis, Online shopping
Chenhao Ma, Shuxian Song, Shun Xu, Shangrui Xiao. Research on Online Shopping Comments Based on Text Emotional Comments. Frontiers in Educational Research (2020) Vol. 3 Issue 7: 68-71. https://doi.org/10.25236/FER.2020.030720.
[1] Lutfullaeva, Malika, Marina Medvedeva, Eugene Komotskiy (2018). “Optimization of Sentiment Analysis Methods for classifying text comments of bank customers.” IFAC-PapersOnLine, vol.51, no.32, pp. 55-60.
[2] Xiaoyun Wang, Lingling Shi (2016). Commodity comprehensive scoring model based on sentiment quantification of online reviews. Journal of Hangzhou Dianzi University (Social Science Edition), vol.12, no.3, pp.8-15.
[3] Wang Longge, Wang Min (2019). A quantitative emotional scoring model based on user comments. Digital Technology and applications, vol.6, pp.71-72.