基于TDMA數(shù)據(jù)鏈的文本分類系統(tǒng)研究與應用
發(fā)布時間：2019-02-13 03:04
【摘要】：無線設備由于其操作簡單、易于攜帶、交換信息方便等特點在現(xiàn)代通信中有十分重要的地位,被廣泛用于各個領域,是軍方不可或缺的通信手段。所以,無論在軍事戰(zhàn)爭還是災難救援中,無線設備是一種常用的指揮和聯(lián)絡工具。為了更快速、可靠的在無線設備之間傳輸數(shù)據(jù),我們構建了一種基于TDMA(Time Division Multiple Access,時分多址)數(shù)據(jù)鏈的無線網(wǎng)絡系統(tǒng)。本文的研究內(nèi)容主要包括該系統(tǒng)中業(yè)務平臺和無線平臺之間的協(xié)議設計和實現(xiàn)以及業(yè)務平臺上中文文本分類系統(tǒng)的研究和實現(xiàn),具體如下:1)在基于TDMA數(shù)據(jù)鏈的無線網(wǎng)絡系統(tǒng)中,每種類型的終端均包含無線平臺和業(yè)務平臺,無線平臺完成各終端信息的傳輸,業(yè)務平臺負責處理文件、定位信息、音頻和視頻等業(yè)務的管理和傳輸,為了保證各個業(yè)務可靠的傳輸,我們設計了 SWIP(Service Wireless Interface Protocol)協(xié)議。2)為了對傳輸內(nèi)容進行過濾和分類處理,在業(yè)務平臺上需要安裝中文文本分類系統(tǒng)。首先,介紹了中文文本分類系統(tǒng),并詳細闡述了系統(tǒng)中各個模塊,如文本預處理、特征降維、文本表示、分類算法等。并提出了新的中文文本分類系統(tǒng),在新的系統(tǒng)中采用LDA(Latent Dirichlet Allocation)主題模型來表示文本,采用支持向量機(Support Vector Machines,SVM)和 K-最鄰近算法(K-nearest neighbor classification)的結合算法KSVM進行分類。該系統(tǒng)使用JAVA編程語言在Windows系統(tǒng)下實現(xiàn)。3)為了分析提出的中文文本分類系統(tǒng)的性能,需要通過實驗計算評價標準例如準確率、召回率、F1-measure等。本文提出了兩種對比實驗:基于LDA主題模型不同分類算法的分類效果分析;基于不同文本模型的KSVM分類效果分析。通過分析實驗數(shù)據(jù)可知:基于LDA主題模型和KSVM分類算法的中文文本分類系統(tǒng)可以獲得更好的分類效果。
[Abstract]:Wireless equipment is widely used in various fields because of its simple operation, easy to carry, convenient exchange of information and so on. It is an indispensable means of communication for the military. Therefore, wireless equipment is a common command and liaison tool in both military war and disaster relief. In order to transmit data between wireless devices more quickly and reliably, we construct a wireless network system based on TDMA (Time Division Multiple Access, (time Division multiple access) data link. The research content of this paper mainly includes the protocol design and implementation between the service platform and wireless platform, and the research and implementation of the Chinese text classification system on the service platform. The details are as follows: 1) in the wireless network system based on TDMA data link, each type of terminal includes wireless platform and service platform. The wireless platform completes the transmission of terminal information, and the business platform is responsible for processing files and locating information. The management and transmission of audio and video services, in order to ensure the reliable transmission of each service, we designed SWIP (Service Wireless Interface Protocol) protocol. 2) in order to filter and classify the transmission content, The Chinese text classification system needs to be installed on the business platform. Firstly, the Chinese text classification system is introduced, and various modules in the system, such as text preprocessing, feature dimensionality reduction, text representation, classification algorithm and so on, are described in detail. A new Chinese text classification system is proposed, in which the LDA (Latent Dirichlet Allocation) topic model is used to represent the text and the support vector machine (Support Vector Machines, is used to represent the text. SVM) and K-nearest neighbor algorithm (K-nearest neighbor classification) combined with the algorithm KSVM to classify. In order to analyze the performance of the proposed Chinese text classification system, the evaluation criteria such as accuracy, recall rate, F1-measure and so on need to be calculated experimentally. This paper presents two comparative experiments: the analysis of classification effect based on different classification algorithms based on LDA subject model and the analysis of KSVM classification effect based on different text models. By analyzing the experimental data, we can see that the Chinese text classification system based on LDA topic model and KSVM classification algorithm can achieve better classification effect.
【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TN919.2;TP391.1
本文編號：2421104
[Abstract]:Wireless equipment is widely used in various fields because of its simple operation, easy to carry, convenient exchange of information and so on. It is an indispensable means of communication for the military. Therefore, wireless equipment is a common command and liaison tool in both military war and disaster relief. In order to transmit data between wireless devices more quickly and reliably, we construct a wireless network system based on TDMA (Time Division Multiple Access, (time Division multiple access) data link. The research content of this paper mainly includes the protocol design and implementation between the service platform and wireless platform, and the research and implementation of the Chinese text classification system on the service platform. The details are as follows: 1) in the wireless network system based on TDMA data link, each type of terminal includes wireless platform and service platform. The wireless platform completes the transmission of terminal information, and the business platform is responsible for processing files and locating information. The management and transmission of audio and video services, in order to ensure the reliable transmission of each service, we designed SWIP (Service Wireless Interface Protocol) protocol. 2) in order to filter and classify the transmission content, The Chinese text classification system needs to be installed on the business platform. Firstly, the Chinese text classification system is introduced, and various modules in the system, such as text preprocessing, feature dimensionality reduction, text representation, classification algorithm and so on, are described in detail. A new Chinese text classification system is proposed, in which the LDA (Latent Dirichlet Allocation) topic model is used to represent the text and the support vector machine (Support Vector Machines, is used to represent the text. SVM) and K-nearest neighbor algorithm (K-nearest neighbor classification) combined with the algorithm KSVM to classify. In order to analyze the performance of the proposed Chinese text classification system, the evaluation criteria such as accuracy, recall rate, F1-measure and so on need to be calculated experimentally. This paper presents two comparative experiments: the analysis of classification effect based on different classification algorithms based on LDA subject model and the analysis of KSVM classification effect based on different text models. By analyzing the experimental data, we can see that the Chinese text classification system based on LDA topic model and KSVM classification algorithm can achieve better classification effect.
【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TN919.2;TP391.1
【參考文獻】
相關期刊論文 前10條
1 陳亮;龔儉;;基于卡方統(tǒng)計的應用協(xié)議流量行為特征分析方法[J];軟件學報;2010年11期
2 黃秀麗;王蔚;;一種改進的文本分類特征選擇方法[J];計算機工程與應用;2009年36期
3 孫鐵利;劉延吉;;中文分詞技術的研究現(xiàn)狀與困難[J];信息技術;2009年07期
4 龍樹全;趙正文;唐華;;中文分詞算法概述[J];電腦知識與技術;2009年10期
5 鄒加棋;陳國龍;郭文忠;;基于圖模型的中文文檔分類研究[J];小型微型計算機系統(tǒng);2006年04期
6 申紅;呂寶糧;內(nèi)山將夫;井佐原均;;文本分類的特征提取方法比較與改進[J];計算機仿真;2006年03期
7 周昭濤,卜東波,程學旗;文本的圖表示初探[J];中文信息學報;2005年02期
8 李蓉 ,葉世偉 ,史忠植;SVM-KNN分類器——一種提高SVM分類精度的新方法[J];電子學報;2002年05期
9 揭春雨,劉源,梁南元;漢語自動分詞實用系統(tǒng)CASS的設計和實現(xiàn)[J];中文信息學報;1991年04期
10 徐輝,何克抗,孫波;書面漢語自動分詞專家系統(tǒng)的實現(xiàn)[J];中文信息學報;1991年03期
相關博士學位論文 前1條
1 尚文倩;文本分類及其相關技術研究[D];北京交通大學;2007年
相關碩士學位論文 前3條
1 王國才;樸素貝葉斯分類器的研究與應用[D];重慶交通大學;2010年
2 曹衛(wèi)峰;中文分詞關鍵技術研究[D];南京理工大學;2009年
3 黃志剛;基于貝葉斯的中文垃圾郵件過濾系統(tǒng)的設計與實現(xiàn)[D];電子科技大學;2007年
，本文編號：2421104
本文鏈接：http://www.sikaile.net/kejilunwen/ruanjiangongchenglunwen/2421104.html
最近更新
教材專著
