Spoken dialog summarization system with HAPPINESS/SUFFERING factor recognition

Abstract

This work presents a spoken dialog summarization system with HAPPINESS/SUFFERING factor recognition. The semantic content is compressed and classified by factor categories from spoken dialog. The transcription of automatic speech recognition is then processed through Chinese Knowledge and Information Processing segmentation system. The proposed system also adopts the part-of-speech tags to effectively select and rank the keywords. Finally, the HAPPINESS/SUFFERING factor recognition is done by the proposed point-wise mutual information. Compared with the original method, the performance is improved by applying the significant scores of keywords. The experimental results show that the average precision rate for factor recognition in outside test can reach 73.5% which demonstrates the possibility and potential of the proposed system.

Keywords

spoken dialog summarization / keyword extraction / natural language processing (NLP) / sentiment analysis

References

[1] LeeH, ShiangS R, YehC F, Chen Y N, HuangY , KongS Y, LeeL S. Spoken knowledge organization by semantic structuring and a prototype course lecture system for personalized learning. IEEE/ACM Transactions on Audio, Speech and Language Processing (TASLP), 2014, 22(5), 883–898

[2] HuangC L, WuC H. Spoken document retrieval using multilevel knowledge and semantic verification. IEEE Transactions on Audio, Speech, and Language Processing, 2007, 15(8): 2551–2560

[3] WangJ F, ChenBW, FanW K, Li C H. Emotion-aware assistive system for humanistic care based on the orange computing concept. Applied Computational Intelligence and Soft Computing, 2012

[4] KoolagudiS G, KumarN, RaoK S. Speech emotion recognition using segmental level prosodic analysis. In: Proceedings of IEEE International Conference on Devices and Communication. 2011, 1–5

[5] AhmedT, IslamM, AhmadM. Human emotion modeling based on salient global features of EEG signal. In: Proceedings of IEEE International Conference on Advances in Electrical Engineering. 2013, 246–251

[6] TsaiH C, FanW K, ChenB W, Wang J F, LinP C . A real-time awareness system for happiness expression based on the multilayer histogram of oriented gradients. In: Proceedings of the 4th International Conference on Awareness Science and Technology. 2012, 289–293

[7] ParkJ S, JangG J, KimJ H. Multistage utterance verification for keyword recognition-based online spoken content retrieval. IEEE Transactions on Consumer Electronics, 2012, 58(3), 1000–1005

[8] LiuF, LiuF F, LiuY. A supervised framework for keyword extraction from meeting transcripts. IEEE Transactions on Audio, Speech, and Language Processing, 2011, 19(3), 538–548

[9] LiuF, LiuY. Towards abstractive speech summarization: exploring unsupervised and supervised approaches for spoken utterance compression. IEEE Transactions on Audio, Speech, and Language Processing, 2013, 21(7), 1469–1480

[10] QinY. Applying frequency and location information to keyword extraction in single document. In: Proceedings of the 2nd IEEE International Conference on Cloud Computing and Intelligent Systems. 2012, 1398–1402

[11] WartenaC, Brussee R, SlakhorstW . Keyword extraction using word co-occurrence. In: Proceedings of IEEE Workshops on Database and Expert Systems Applications. 2010, 54–58

[12] JiaoH, LiuQ, JiaH B. Chinese keyword extraction based on n-gram and word co-occurrence. In: Proceedings of IEEE International Conference on Computational Intelligence and Security Workshops. 2007, 152–155

[13] GuptaA, DixitA, SharmaA K. A novel statistical and linguistic features based technique for keyword extraction. In: Proceedings of IEEE International Conference on Information Systems and Computer Networks. 2014, 55–59

[14] HuX H, WuB. Automatic keyword extraction using linguistic features. In: Proceedings of the 6th IEEE International Conference on Data Mining Workshops. 2006, 19–23

[15] ChenB W, RhoS, GuizaniM, Fan W K. Cognitive sensors based on kernel ridge phase-smoothing localization and multiregional histograms of oriented gradients. IEEE Transactions on Emerging Topics in Computing, 2016

[16] TroykaL Q, Thweatt J W. Structured Reading. Englewood Cliffs, NJ: Prentice Hall, 2003

[17] HuangC L, HsiehC H, WuC H. Spoken document summarization using acoustic, prosodic and semantic information. In: Proceedings of IEEE International Conference on Multimedia. 2005

[18] KurmiR, JainP. Text summarization using enhanced MMR technique. In: Proceedings of International Conference on Computer Communication and Informatics. 2014, 1–5

[19] LiY B, Merialdo B. Multi-video summarization based on Video- MMR. In: Proceedings of the 11th International Workshop on Image Analysis for Multimedia Interactive Services, 2010, 1–4

[20] FerreiraR, Freitas F, CabralL S , LinsR D, LimaR, Fran�aG, Simskez S J, FavaroL . A four dimension graph model for automatic text summarization. In: Proceedings of IEEE/WIC/ACM International Joint Conferences onWeb Intelligence (WI) and Intelligent Agent Technologies. 2013, 389–396

[21] CambriaE, Schuller B, XiaY Q , HavasiC. New avenues in opinion mining and sentiment analysis. IEEE Intelligent Systems, 2013, 15–21

[22] OrtonyA, CloreG L, CollinsA. The Cognitive Structure of Emotions. Cambridge: Cambridge University Press, 1990

[23] StevensonR A, MikelsJ A, JamesT W. Characterization of the affective norms for English words by discrete emotional categories. Behavior Research Methods, 2007, 39(4): 1020–1024

[24] GrassiM, Cambria E, HussainA , PiazzaF. Sentic web: a new paradigm for managing social media affective information. Cognitive Computation, 2011, 3(3), 480–489

[25] ChenB W, HeX Y, JiW, RhoS, KungS Y. Support vector analysis of large-scale data based on kernels with iteratively increasing order. Journal of Supercomputing, 2016, 72(9): 3297–3311

[26] OlsherD J. Full spectrum opinion mining: integrating domain, syntactic and lexical knowledge. In: Proceedings of the 12th IEEE International Conference on Data Mining Workshops. 2012, 693–700

[27] WuC H, LiangW B. Emotion recognition of affective speech based on multiple classifiers using acoustic-prosodic information and semantic labels. IEEE Transactions on Affective Computing, 2011, 2(1): 10–21

[28] XieS S, LiuY. Using n-best lists and confusion networks for meeting summarization.IEEE Transactions on Audio, Speech, and Language Processing, 2011, 19(5), 1160–1169

[29] ChenB W, JiW. Intelligent marketing in smart cities: crowdsourced data for geo-conquesting. IT Professional, 2016, 18(4): 18–24

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg