摘要： 利用隐马尔可夫模型(HMM)对多媒体数据仓库进行复杂数据挖掘，复杂数据挖掘要解决的难题是音频和视频识别。建立音频和视频的识别模型及其相关的算法，在视频识别算法上构造出符合HMM的识别方法。根据模型建立系统, 实验证明声音的识别率最高达到96.67%, 视频中特征值的检测率可达87.81%。研究结果可以应用在多媒体的识别和数据挖掘领域，提供一个比较完整的复杂数据挖掘的模型和算法。

关键词: HMM, 复杂数据挖掘, 音频识别, 视频检测

Abstract: This paper mainly discusses a method to process complex data mining in the multimedia data ware by HMM (Hidden Markov Model) . The problem of complex data mining is the recognition of the audio and video. This paper builds the model of recognition of audio and video. It constructs the method according to the HMM. Due to results of experiments, the highest ratio of recognition of audio is 96.67%, the highest ratio of recognition of video is 87.81%. The application area of the results covers recognition of multimedia and the fields of data mining. This paper provides complete models and technique of CDM (Complex Data Mining).

中图分类号: 

• TP311.13