面向大数据的语音日志分析系统的设计与实现

发布时间：2018-09-18 19:56

【摘要】：大数据时代已经来临并已有阶段性的发展,随着基础设备进一步完善,越来越多的数据得以被发觉、利用。许多公司甚至提出了无数据,不智能的口号。市场上也涌现出了许多商业智能公司的产品,催生出了数据分析产业这一跨学科的边缘科学行业,整个数据分析产业链包括数据采集、数据存储、数据处理、数据分析及结果呈现,商业决策。目前,市场上的数据分析产品基本停留在对数据的统计,分析指标不够细化的阶段,本文围绕用户使用产品产生的语音数据进行分析,旨在搭建一个功能完善,交互性强,可扩展的语音日志分析系统,这对推进数据分析产业产品的研发具有重大意义。为搭建完整的语音日志分析系统,本文参考大量文献资料与市场产品,分析了该系统所需技术,数据分析所需方法,学习研究语音日志分析系统关键技术,主要研究工作包括:第一,为解决如何展示语音日志分析结果的问题,提出一种基于Django框架的语音日志分析系统搭建方法。Django是由Python编写,基于MVC架构的一个开源的web框架。开发者可根据此框架快速开发自己的web应用。本文完成了基于Django框架的语音日志分析系统的搭建,使用HTML、CSS、JS语言编写页面,为语音数据分析结果的展示奠定了基础。第二,分别从统计数据分析和数据挖掘分析两个方面对语音数据进行了全面分析。统计数据分析包括数据分析平台常见的分析指标如服务量,用户量,人均服务量及增长率,机型排名,地域排名等。数据挖掘分析包括以数据挖掘技术、语音纠错技术等为核心的数据分析技术,完成了用户分群、流失用户预测、用户画像等功能。第三,对语音日志分析系统的功能进行了扩展。结合用户语音与影视数据,在对真实数据的充分分析的前提下,提出了一种结合用户画像的相关电影推荐方法。经证明,该方法具有较高的推荐效果。
[Abstract]:Big data era has come and has been phased development, with the further improvement of infrastructure, more and more data can be found and used. Many companies even put forward the slogan of no data, no intelligence. There are also many products of business intelligence companies emerging in the market, resulting in the data analysis industry, an interdisciplinary and fringe scientific industry. The entire data analysis industry chain includes data collection, data storage, data processing, Data analysis and presentation of results, business decisions. At present, the data analysis products in the market basically stay at the stage of statistics of the data, the analysis index is not detailed enough. This paper focuses on the analysis of the voice data generated by the user using the product, in order to build a perfect function and strong interactivity. The scalable voice log analysis system is of great significance to the R & D of data analysis industry. In order to build a complete voice log analysis system, this paper, referring to a large amount of literature and market products, analyzes the required technology and methods of data analysis, and studies the key technologies of the voice log analysis system. The main research work includes: first, in order to solve the problem of how to display the results of voice log analysis, this paper proposes a method to build a voice log analysis system based on Django framework. Django is an open source web framework written by Python and based on MVC architecture. Developers can quickly develop their own web applications based on this framework. In this paper, a voice log analysis system based on Django framework is built, and the page is written in HTML,CSS,JS language, which lays a foundation for the display of voice data analysis results. Secondly, the speech data are analyzed from two aspects: statistical data analysis and data mining analysis. Statistical data analysis includes common analysis indicators such as service volume, user quantity, per capita service volume and growth rate, aircraft ranking, regional ranking and so on. Data mining analysis includes data analysis technology based on data mining technology, voice correction technology and so on. The functions of user grouping, losing user prediction, user portrait and so on are accomplished. Thirdly, the function of voice log analysis system is extended. Based on the analysis of the real data and the user voice and video data, this paper proposes a new method to recommend the relevant films with the help of the user's portrait. It has been proved that this method has high recommended effect.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【相似文献】