Fusion effect of SVM in spark architecture for speech data mining in cluster structure
- PDF / 1,324,856 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 81 Downloads / 181 Views
Fusion effect of SVM in spark architecture for speech data mining in cluster structure Jianfei Shen1 · Harry Haoxiang Wang2 Received: 30 January 2020 / Accepted: 27 April 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Fusion effect of SVM in the Spark architecture for speech data mining in cluster structure is studied in this manuscript. Based on the information entropy of nodes, the data in clusters are fused to eliminate redundant data and improve the efficiency of information fusion. Information entropy is a statistical form based on the characteristics of information representation, which reflects the average amount of information in information. Based on the Spark platform SVM algorithm, the frequent items with the highest support after each sort are directly recursively obtained, and the transaction data set is allocated to each computing node. The structure of the item head table directly affects the efficiency of the algorithm, so optimizing the structure of the item head table can improve the efficiency of the algorithm in constructing FP-Tree, and then improve the efficiency of the whole algorithm. The proposed speech data mining algorithm can cluster, analyze, and comprehensively detection the saliency information, the detection accuracy is much higher than the state-of-the-art models. The experimental results compared with the latest research have reflected that fact that the proposed model has the better performance and robustness. Keywords SVM · Spark architecture · Data mining · Speech analysis · Cluster structure · Fusion effect
1 Introduction In daily communication of human beings, language communication is required, and the most important way of language communication is voice communication. The voice signal contains a lot of information, and it also has the very high variability, including changes in tone and mood, etc. Therefore, the study of this unique change can very effectively improve the success rate of artificial intelligence speech emotion recognition. In the field of artificial intelligence, in the process of human–computer communication, speech has a very high amount of information for the information transmission, fast and practical, so that the research on speech emotion recognition method has gradually become a hot research point in the near future. The technology of speech and emotion recognition is mainly through receiving the voice sent by people, and processing and judging
* Jianfei Shen [email protected] 1
Visual Arts Department, Hunan Mass Media Vocational Technical College, Changsha, Hunan, China
GoPerception Laboratory, Ithaca, USA
2
these speech, analyzing the real intention that people want to express, so that they can carry on the human–computer interaction normally and reasonably. Although the current AI speech recognition technology has made great progress, AI’s recognition of sound languages still needs to be trained and optimized with large amounts of the data according to the application scenario. Each language
Data Loading...
