一种自适应的情感灰色预测模型
|
摘要
目前,一些研究工作已使用评论中蕴含的情感信息对产品的销量进行预测,这些预测方法大部分偏重于评论的整体情感或情绪,忽略了产品特征的情感信息.针对这一问题,本文提出了一种自适应的情感灰色预测模型(Adaptive Sentiment Grey Prediction M odel,ASGPM).在预测时,首先通过条件随机场模型建立产品特征库,并量化情感词典中情感词的情感强度;然后从评论中计算产品多个特征的情感强度,每个特征的情感强度分别与产品销量建立自适应的灰色模型(Adaptive Multivariable Grey M odel,AM GM),并进行销量预测;最后,将产品销量与多个销量预测结果通过ASGPM模型进行预测.实验结果表明,该预测方法销量的动态预测方法中具有较好的预测性能.
Now some researches exploit the sentiment of reviews to predict product sales and obtain some achievements. However,the existing prediction methods focus on the document-level sentiment or emotion of the review,and fail to consider feature sentiment in the review. To cope with this problem,an approach called ASGPM( Adaptive Sentiment Grey Prediction Model) is proposed in this paper. In the prediction process,a feature library is constructed with conditional random fields model and each sentiment word in the sentiment dictionary is quantified at first; Then the intensity of each feature sentiment is gained from the reviews,a single model called AMGM( Adaptive Multivariate Grey M odel,AMGM) is constructed by combining the intensity of each feature sentiment with historical product sales respectively,and predict the future sales via the model; Finally,the historical product sales combined with all predict results is used to forecast the product sales via ASGPM. Experiment results show that the proposed approach outperforms several stateof-the-art forecast methods in the dynamical prediction.
Now some researches exploit the sentiment of reviews to predict product sales and obtain some achievements. However,the existing prediction methods focus on the document-level sentiment or emotion of the review,and fail to consider feature sentiment in the review. To cope with this problem,an approach called ASGPM( Adaptive Sentiment Grey Prediction Model) is proposed in this paper. In the prediction process,a feature library is constructed with conditional random fields model and each sentiment word in the sentiment dictionary is quantified at first; Then the intensity of each feature sentiment is gained from the reviews,a single model called AMGM( Adaptive Multivariate Grey M odel,AMGM) is constructed by combining the intensity of each feature sentiment with historical product sales respectively,and predict the future sales via the model; Finally,the historical product sales combined with all predict results is used to forecast the product sales via ASGPM. Experiment results show that the proposed approach outperforms several stateof-the-art forecast methods in the dynamical prediction.
引文
[1]Liu Y,Huang X,An A,et al.ARSA:asentiment-aware model for predicting sales performance using blogs[C].Proceedings of the30th ACM SIGIR International Conference on Research and Development in Information Retrieval(SIGIR'07),Amsterdam,USA,2007:607-614.
[2]Ye Q,Li H,Wang Z,et al.The influence of hotel price on perceived service quality and value in e-tourism an empirical investigation based on online traveler review s[J].Journal of Hospitality&Tourism Research,2014,38(1):23-39.
[3]Du J,Xu H,Huang X.Box office prediction based on microblog[J].Expert Systems with Applications,2014,41(4):1680-1689.
[4]Yu X,Liu Y,Huang X,et al.Mining online reviews for predicting sales performance:a case study in the movie domain[J].IEEE Transactions on Know ledge and Data Engineering,2012,24(4):720-734.
[5]Chong A Y L,Li B,Ngai E W T,et al.Predicting online product sales via online review s,sentiments,and promotion strategies:a big data architecture and neural netw ork approach[J].International Journal of Operations&Production M anagement,2016,36(4):358-383.
[6]Li Xue-ni,Zhang Shao-wu,Yang Liang,et al.ARES:autoregressive emotion-sensitive model for predicting sales performance[J].Journal of Computer Research and Development,2013,50(8):1722-1727.
[7]Ahn H I,Spangler W S.Sales prediction with social media analysis[C].Proceedings of 2014 Annual SRII Global Conference(SRII'14),San Jose,USA,2014:213-222.
[8]Gui Bin,Huang Li-dong,Zhou Jie,et al.Research on forecasting the fuzzy boundary of financial time series[J].Journal of Chinese Computer Systems,2012,33(10):2283-2286.
[9]Liu Si-feng.Grey system theory and its application[M].Beijing:Science Press,2010.
[10]Wang Ya-nan,Lei Ying-jie,Lei Yang,et al.High order intuitionistic fuzzy time series forecasting model[J].Journal on Communications,2016,37(5):115-124.
[11]Wei Lin,Zhou Lei,Wu Ning,et al.A predictive thermal model combined w ith the second derivative for multiprocessor system-onchips[J].Acta Electronica Sinica,2016,44(6):1272-1278.
[12]Han Min,Xu Mei-ling,Ren Wei-jie.Research on multivariate chaotic time series prediction using mRSM model[J].Acta Automatica Sinica,2014,40(5):822-829.
[13]Bezuglov A,Comert G.Short-term freeway traffic parameter prediction:application of grey system theory models[J].Expert Systems w ith Applications,2016,62(1):284-292.
[14]He Z,Shen Y,Li J,et al.Regularized multivariable grey model for stable grey coefficients estimation[J].Expert Systems w ith Applications,2015,42(4):1806-1815.
[15]Ma Jun,Liu Xiao-dong,Meng Ying.Research of urban traffic flow forecasting based on neural netw ork[J].Acta Electronica Sinica,2009,37(5):1092-1094.
[16]Floyd K,Freling R,Alhoqail S,et al.How online product reviews affect retail sales:A meta-analysis[J].Journal of Retailing,2014,90(2):217-232.
[6]李雪妮,张绍武,杨亮,等.ARES:用于预测的情感感知自回归模型[J].计算机研究与发展,2013,50(8):1722-1727.
[8]桂斌,黄立冬,周杰,等.金融时间序列模糊边界预测研究[J].小型微型计算机系统,2012,33(10):2283-2286.
[9]刘思峰.灰色系统理论及其应用[M].北京:科学出版社,2010.
[10]王亚男,雷英杰,雷阳,等.高阶直觉模糊时间序列预测模型[J].通信学报,2016,37(5):115-124.
[11]魏琳,周磊,吴宁,等.多处理器片上系统中一种结合二阶导数的温度预测模型[J].电子学报,2016,44(6):1272-1278.
[12]韩敏,许美玲,任伟杰.多元混沌时间序列的相关状态机预测模型研究[J].自动化学报,2014,40(5):822-829.
[15]马君,刘小冬,孟颖.基于神经网络的城市交通流预测研究[J].电子学报,2009,37(5):1092-1094.
[2]Ye Q,Li H,Wang Z,et al.The influence of hotel price on perceived service quality and value in e-tourism an empirical investigation based on online traveler review s[J].Journal of Hospitality&Tourism Research,2014,38(1):23-39.
[3]Du J,Xu H,Huang X.Box office prediction based on microblog[J].Expert Systems with Applications,2014,41(4):1680-1689.
[4]Yu X,Liu Y,Huang X,et al.Mining online reviews for predicting sales performance:a case study in the movie domain[J].IEEE Transactions on Know ledge and Data Engineering,2012,24(4):720-734.
[5]Chong A Y L,Li B,Ngai E W T,et al.Predicting online product sales via online review s,sentiments,and promotion strategies:a big data architecture and neural netw ork approach[J].International Journal of Operations&Production M anagement,2016,36(4):358-383.
[6]Li Xue-ni,Zhang Shao-wu,Yang Liang,et al.ARES:autoregressive emotion-sensitive model for predicting sales performance[J].Journal of Computer Research and Development,2013,50(8):1722-1727.
[7]Ahn H I,Spangler W S.Sales prediction with social media analysis[C].Proceedings of 2014 Annual SRII Global Conference(SRII'14),San Jose,USA,2014:213-222.
[8]Gui Bin,Huang Li-dong,Zhou Jie,et al.Research on forecasting the fuzzy boundary of financial time series[J].Journal of Chinese Computer Systems,2012,33(10):2283-2286.
[9]Liu Si-feng.Grey system theory and its application[M].Beijing:Science Press,2010.
[10]Wang Ya-nan,Lei Ying-jie,Lei Yang,et al.High order intuitionistic fuzzy time series forecasting model[J].Journal on Communications,2016,37(5):115-124.
[11]Wei Lin,Zhou Lei,Wu Ning,et al.A predictive thermal model combined w ith the second derivative for multiprocessor system-onchips[J].Acta Electronica Sinica,2016,44(6):1272-1278.
[12]Han Min,Xu Mei-ling,Ren Wei-jie.Research on multivariate chaotic time series prediction using mRSM model[J].Acta Automatica Sinica,2014,40(5):822-829.
[13]Bezuglov A,Comert G.Short-term freeway traffic parameter prediction:application of grey system theory models[J].Expert Systems w ith Applications,2016,62(1):284-292.
[14]He Z,Shen Y,Li J,et al.Regularized multivariable grey model for stable grey coefficients estimation[J].Expert Systems w ith Applications,2015,42(4):1806-1815.
[15]Ma Jun,Liu Xiao-dong,Meng Ying.Research of urban traffic flow forecasting based on neural netw ork[J].Acta Electronica Sinica,2009,37(5):1092-1094.
[16]Floyd K,Freling R,Alhoqail S,et al.How online product reviews affect retail sales:A meta-analysis[J].Journal of Retailing,2014,90(2):217-232.
[6]李雪妮,张绍武,杨亮,等.ARES:用于预测的情感感知自回归模型[J].计算机研究与发展,2013,50(8):1722-1727.
[8]桂斌,黄立冬,周杰,等.金融时间序列模糊边界预测研究[J].小型微型计算机系统,2012,33(10):2283-2286.
[9]刘思峰.灰色系统理论及其应用[M].北京:科学出版社,2010.
[10]王亚男,雷英杰,雷阳,等.高阶直觉模糊时间序列预测模型[J].通信学报,2016,37(5):115-124.
[11]魏琳,周磊,吴宁,等.多处理器片上系统中一种结合二阶导数的温度预测模型[J].电子学报,2016,44(6):1272-1278.
[12]韩敏,许美玲,任伟杰.多元混沌时间序列的相关状态机预测模型研究[J].自动化学报,2014,40(5):822-829.
[15]马君,刘小冬,孟颖.基于神经网络的城市交通流预测研究[J].电子学报,2009,37(5):1092-1094.