本文選題：網(wǎng)絡(luò)搜索　切入點：投資者情緒　出處：《廈門大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：大數(shù)據(jù)信息和股票改革是近年來經(jīng)濟(jì)發(fā)展熱點,探索大數(shù)據(jù)信息作用于股票市場投資領(lǐng)域的機制就具有十分重要的現(xiàn)實意義。國內(nèi)外的理論研究和實踐證明,投資者情緒對股票市場未來趨勢具有一定的預(yù)測作用與溢出效應(yīng),但對于利用大數(shù)據(jù)網(wǎng)絡(luò)搜索信息構(gòu)造股票市場先行情緒指標(biāo)的研究仍然存在空白；我國作為世界第二經(jīng)濟(jì)大體,擁有上海證券交易所與深圳證券交易所兩個證券交易市場,它們是我國資本流通的主要市場,是經(jīng)濟(jì)發(fā)展的主要風(fēng)向標(biāo)。因此基于網(wǎng)絡(luò)搜索信息構(gòu)造適應(yīng)我國股票市場的投資者情緒指數(shù),并探究投資者情緒與股票市場之間的相互溢出效應(yīng)就具有一定的理論和研究價值。 本文綜合運用統(tǒng)計學(xué)、時間序列模型、金融學(xué)等理論方法,借鑒國內(nèi)外關(guān)于投資者情緒與股票市場的研究成果和實踐經(jīng)驗,對我國滬深投資者情緒與股票市場進(jìn)行多維度的研究。本文在深入理解投資者情緒與股票市場內(nèi)涵的基礎(chǔ)上,首先利用文本挖掘技術(shù)構(gòu)造以“股票市場”為核心的谷歌網(wǎng)絡(luò)搜索初始詞庫,隨后利用時差相關(guān)系數(shù)法、隨機森林、CART、神經(jīng)網(wǎng)絡(luò)等算法對260多個候選關(guān)鍵詞進(jìn)行信息約簡,并利用主成分和隨機森林重要性方法分別構(gòu)造了具有先行預(yù)測作用的滬深兩市投資者情緒指數(shù)；其次利用VAR(n)-BEKK(1,1)-GARCH模型分別探討了滬深投資者情緒與股票市場指數(shù)、滬深投資者情緒等三種關(guān)系之間的溢出效應(yīng)機制,最后基于它們之間的溢出效應(yīng)關(guān)系對投資者情緒與股票市場進(jìn)行綜合評價。主要結(jié)論有：一是隨機森林算法具有較好的變量篩選作用；二是社會文本信息與網(wǎng)絡(luò)數(shù)據(jù)對推斷我國投資者情緒具有良好的信息參考價值；三是基于谷歌網(wǎng)絡(luò)搜索數(shù)據(jù)所構(gòu)建的投資者情緒對滬深兩市具有良好的預(yù)測作用；四是投資者情緒在滬深兩市、不同時期存在差異；五是滬深兩市的投資情緒與市場指數(shù)之間存在聯(lián)動機制；六是滬深兩市的投資者情緒與市場指數(shù)都存在較強烈的波動反自身性；七是滬深兩市的投資者情緒與市場指數(shù)分別存在雙向溢出效應(yīng)；八是滬深兩市投資者情緒存在跨市溢出效應(yīng)。
[Abstract]:Big data's information and stock reform has been a hot spot of economic development in recent years, so it is of great practical significance to explore the mechanism of big data's information acting on the field of investment in the stock market. Investor sentiment has certain predictive effect and spillover effect on the future trend of stock market, but there is still a blank in the research of constructing stock market leading emotion index by using big data network search information. As the second largest economy in the world, China has two securities trading markets, the Shanghai Stock Exchange and the Shenzhen Stock Exchange, which are the main markets for the circulation of capital in China. It is the main vane of economic development, so it has certain theoretical and research value to construct the investor sentiment index based on the network search information to adapt to the stock market of our country, and to explore the mutual spillover effect between investor sentiment and stock market. This paper synthetically uses statistics, time series model, finance and other theoretical methods to learn from the research results and practical experience of investor sentiment and stock market at home and abroad. Based on the deep understanding of investor sentiment and the connotation of stock market, this paper makes a multi-dimensional study on investor sentiment and stock market in Shanghai and Shenzhen. First of all, using text mining technology to construct the initial lexicon of Google Web search based on "stock market", then using the method of time-difference correlation coefficient, stochastic forest cart, neural network and other algorithms to reduce the information of more than 260 candidate keywords. Using the principal component and the stochastic forest importance method, the investor sentiment index of Shanghai and Shenzhen stock markets is constructed, and the investor sentiment index and stock market index of Shanghai and Shenzhen stock markets are discussed by using the VARN model and the VARN BEKKGARCH model, respectively. The spillover effect mechanism between the three relationships, such as investor sentiment in Shanghai and Shenzhen, Finally, the investors' sentiment and stock market are evaluated based on their spillover effects. The main conclusions are as follows: first, the stochastic forest algorithm has better variable screening function; Second, the social text information and the network data have the good information reference value to infer our country investor sentiment, third, the investor sentiment based on the Google network search data has the good forecast function to the Shanghai and Shenzhen stock market; Fourth, investor sentiment is different in Shanghai and Shenzhen stock markets in different periods, and fifth, there is a linkage mechanism between investment sentiment and market index in Shanghai and Shenzhen stock markets. Sixth, investor sentiment and market index of Shanghai and Shenzhen stock market have strong fluctuation and anti-self, seventh, investor sentiment and market index of Shanghai and Shenzhen stock market have two-way spillover effect respectively. Eight is Shanghai and Shenzhen stock market investor sentiment exists cross-market spillover effect.
【學(xué)位授予單位】：廈門大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：F224;F832.51

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 蔡毅;邢巖;胡丹;;敏感性分析綜述[J];北京師范大學(xué)學(xué)報(自然科學(xué)版);2008年01期

2 馬進(jìn);關(guān)偉;;我國股票市場與宏觀經(jīng)濟(jì)關(guān)系的實證分析[J];財經(jīng)問題研究;2006年08期

3 張俊生,盧賢義,楊熠;噪聲理論能解釋我國封閉式基金折價交易現(xiàn)象嗎——與薛剛、顧鋒、黃培清三位先生商榷[J];財經(jīng)研究;2001年05期

4 徐曉光;黃國輝;;我國股價波動的政策干預(yù)效應(yīng)——基于ARCH類修正模型的實證分析[J];當(dāng)代經(jīng)濟(jì)研究;2007年04期

5 陸安生,陳永強,屠浩文;決策樹C5算法的分析與應(yīng)用[J];電腦知識與技術(shù);2005年09期

6 韓立巖;伍燕然;;投資者情緒與IPOs之謎——抑價或者溢價[J];管理世界;2007年03期

7 池麗旭;張廣勝;莊新田;宋大雷;;投資者情緒指標(biāo)與股票市場——基于擴(kuò)展卡爾曼濾波方法的研究[J];管理工程學(xué)報;2012年03期

8 張雅慧;萬迪f ;付雷鳴;;基于投資者關(guān)注的媒體報道影響投資行為的實驗研究[J];系統(tǒng)工程;2012年10期

9 彭文平,肖繼輝;股市政策與股市波動[J];上海經(jīng)濟(jì)研究;2002年03期

10 唐任伍;倪馨;向守方;;強外生性股票市場政府與交易主體行為博弈[J];經(jīng)濟(jì)管理;2007年10期

，

本文編號：1590754