【摘要】：無(wú)線設(shè)備由于其操作簡(jiǎn)單、易于攜帶、交換信息方便等特點(diǎn)在現(xiàn)代通信中有十分重要的地位,被廣泛用于各個(gè)領(lǐng)域,是軍方不可或缺的通信手段。所以,無(wú)論在軍事戰(zhàn)爭(zhēng)還是災(zāi)難救援中,無(wú)線設(shè)備是一種常用的指揮和聯(lián)絡(luò)工具。為了更快速、可靠的在無(wú)線設(shè)備之間傳輸數(shù)據(jù),我們構(gòu)建了一種基于TDMA(Time Division Multiple Access,時(shí)分多址)數(shù)據(jù)鏈的無(wú)線網(wǎng)絡(luò)系統(tǒng)。本文的研究?jī)?nèi)容主要包括該系統(tǒng)中業(yè)務(wù)平臺(tái)和無(wú)線平臺(tái)之間的協(xié)議設(shè)計(jì)和實(shí)現(xiàn)以及業(yè)務(wù)平臺(tái)上中文文本分類系統(tǒng)的研究和實(shí)現(xiàn),具體如下:1)在基于TDMA數(shù)據(jù)鏈的無(wú)線網(wǎng)絡(luò)系統(tǒng)中,每種類型的終端均包含無(wú)線平臺(tái)和業(yè)務(wù)平臺(tái),無(wú)線平臺(tái)完成各終端信息的傳輸,業(yè)務(wù)平臺(tái)負(fù)責(zé)處理文件、定位信息、音頻和視頻等業(yè)務(wù)的管理和傳輸,為了保證各個(gè)業(yè)務(wù)可靠的傳輸,我們?cè)O(shè)計(jì)了 SWIP(Service Wireless Interface Protocol)協(xié)議。2)為了對(duì)傳輸內(nèi)容進(jìn)行過(guò)濾和分類處理,在業(yè)務(wù)平臺(tái)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(xì)闡述了系統(tǒng)中各個(gè)模塊,如文本預(yù)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(lái)表示文本,采用支持向量機(jī)(Support Vector Machines,SVM)和 K-最鄰近算法(K-nearest neighbor classification)的結(jié)合算法KSVM進(jìn)行分類。該系統(tǒng)使用JAVA編程語(yǔ)言在Windows系統(tǒng)下實(shí)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(guò)實(shí)驗(yàn)計(jì)算評(píng)價(jià)標(biāo)準(zhǔn)例如準(zhǔn)確率、召回率、F1-measure等。本文提出了兩種對(duì)比實(shí)驗(yàn):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(guò)分析實(shí)驗(yàn)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(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TN919.2;TP391.1

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 陳亮;龔儉;;基于卡方統(tǒng)計(jì)的應(yīng)用協(xié)議流量行為特征分析方法[J];軟件學(xué)報(bào);2010年11期

2 黃秀麗;王蔚;;一種改進(jìn)的文本分類特征選擇方法[J];計(jì)算機(jī)工程與應(yīng)用;2009年36期

3 孫鐵利;劉延吉;;中文分詞技術(shù)的研究現(xiàn)狀與困難[J];信息技術(shù);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(shí)與技術(shù);2009年10期

5 鄒加棋;陳國(guó)龍;郭文忠;;基于圖模型的中文文檔分類研究[J];小型微型計(jì)算機(jī)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(jìn)[J];計(jì)算機(jī)仿真;2006年03期

7 周昭濤,卜東波,程學(xué)旗;文本的圖表示初探[J];中文信息學(xué)報(bào);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(xué)報(bào);2002年05期

9 揭春雨,劉源,梁南元;漢語(yǔ)自動(dòng)分詞實(shí)用系統(tǒng)CASS的設(shè)計(jì)和實(shí)現(xiàn)[J];中文信息學(xué)報(bào);1991年04期

10 徐輝,何克抗,孫波;書面漢語(yǔ)自動(dòng)分詞專家系統(tǒng)的實(shí)現(xiàn)[J];中文信息學(xué)報(bào);1991年03期

相關(guān)博士學(xué)位論文 前1條

1 尚文倩;文本分類及其相關(guān)技術(shù)研究[D];北京交通大學(xué);2007年

相關(guān)碩士學(xué)位論文 前3條

1 王國(guó)才;樸素貝葉斯分類器的研究與應(yīng)用[D];重慶交通大學(xué);2010年

2 曹衛(wèi)峰;中文分詞關(guān)鍵技術(shù)研究[D];南京理工大學(xué);2009年

3 黃志剛;基于貝葉斯的中文垃圾郵件過(guò)濾系統(tǒng)的設(shè)計(jì)與實(shí)現(xiàn)[D];電子科技大學(xué);2007年

，

本文編號(hào)：2421104