東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Application of Deep Neural Network M...

Zhang, Liang.

FindBook

Google Book

Amazon

博客來

Application of Deep Neural Network Models to Financial Opinion Mining.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Application of Deep Neural Network Models to Financial Opinion Mining./
作者:	Zhang, Liang.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2018,
面頁冊數:	83 p.
附註:	Source: Dissertations Abstracts International, Volume: 80-05, Section: B.
Contained By:	Dissertations Abstracts International80-05B.
標題:	Finance. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10930862
ISBN:	9780438585850

Application of Deep Neural Network Models to Financial Opinion Mining.
Zhang, Liang.

Application of Deep Neural Network Models to Financial Opinion Mining. - Ann Arbor : ProQuest Dissertations & Theses, 2018 - 83 p.

Source: Dissertations Abstracts International, Volume: 80-05, Section: B.

Thesis (Ph.D.)--State University of New York at Stony Brook, 2018.

This item must not be sold to any third party vendors.

Domain-specific sentiment analysis has obtained an increasing attention from the Natural Language Processing community. However, we find a lack of studies seeking to identify actual opinions in financial articles and comments with bullish or bearish investment recommendations. In our work, we focus on applying neural network models to sentiment analysis task in financial domain. It mainly consists of three parts: (1) we develop an effective lexicon to accurately differ positive and negative sentiments of investors that can help direct financial comments to their actual meanings. Specifically, we firstly apply an efficient prediction-based neural network model to generate initial domain-specific lexicons for short financial texts, such as bullish and bearish comments in tweets, blogs and news headlines. Then, an embedding subspace projection method with 1-vs-rest final layer of sigmoids is utilized to further reduce the unwanted neutral expressions in the induced lexicons; (2) we propose a context-aware deep embedding network for investment recommendation tweets sentiment analysis, by jointly encoding the global and local contextual information. Especially, we capture and include an attitude-aware user embedding to enhance the performance of our model; (3) we identify long and short investment opinions from well-structured summary text of investment recommendation blogs. Current methods tend to regard it as a general text categorization task without considering the inherent content and linguistic specialties. By contrast, we propose to design an architecture that fully utilizes aspects and temporal references knowledge. More specifically, we extend a hierarchical attention neural network by adding important words supervision, which enables the model to pay more attention to sharp words and nearby contexts. In particular, aspects and temporal references position-aware attention mechanism is utilized to achieve focal attention. Experiments on financial lexicons, tweets, and article summary, show that our proposed architectures outperform the state-of-the-art.

ISBN: 9780438585850Subjects--Topical Terms:

542899
Finance.

Application of Deep Neural Network Models to Financial Opinion Mining.
LDR:03184nmm a2200325 4500 001 2207828
005 20190923114238.5
008 201008s2018 ||||||||||||||||| ||eng d
020 $a 9780438585850
035 $a (MiAaPQ)AAI10930862
035 $a (MiAaPQ)grad.sunysb:13862
035 $a AAI10930862
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zhang, Liang. $3 1613704
245 1 0 $a Application of Deep Neural Network Models to Financial Opinion Mining.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2018
300 $a 83 p.
500 $a Source: Dissertations Abstracts International, Volume: 80-05, Section: B.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Li, Xiaolin;Wu, Song.
502 $a Thesis (Ph.D.)--State University of New York at Stony Brook, 2018.
506 $a This item must not be sold to any third party vendors.
520 $a Domain-specific sentiment analysis has obtained an increasing attention from the Natural Language Processing community. However, we find a lack of studies seeking to identify actual opinions in financial articles and comments with bullish or bearish investment recommendations. In our work, we focus on applying neural network models to sentiment analysis task in financial domain. It mainly consists of three parts: (1) we develop an effective lexicon to accurately differ positive and negative sentiments of investors that can help direct financial comments to their actual meanings. Specifically, we firstly apply an efficient prediction-based neural network model to generate initial domain-specific lexicons for short financial texts, such as bullish and bearish comments in tweets, blogs and news headlines. Then, an embedding subspace projection method with 1-vs-rest final layer of sigmoids is utilized to further reduce the unwanted neutral expressions in the induced lexicons; (2) we propose a context-aware deep embedding network for investment recommendation tweets sentiment analysis, by jointly encoding the global and local contextual information. Especially, we capture and include an attitude-aware user embedding to enhance the performance of our model; (3) we identify long and short investment opinions from well-structured summary text of investment recommendation blogs. Current methods tend to regard it as a general text categorization task without considering the inherent content and linguistic specialties. By contrast, we propose to design an architecture that fully utilizes aspects and temporal references knowledge. More specifically, we extend a hierarchical attention neural network by adding important words supervision, which enables the model to pay more attention to sharp words and nearby contexts. In particular, aspects and temporal references position-aware attention mechanism is utilized to achieve focal attention. Experiments on financial lexicons, tweets, and article summary, show that our proposed architectures outperform the state-of-the-art.
590 $a School code: 0771.
650 4 $a Finance. $3 542899
650 4 $a Computer science. $3 523869
690 $a 0508
690 $a 0984
710 2 $a State University of New York at Stony Brook. $b Applied Mathematics and Statistics. $3 1684041
773 0 $t Dissertations Abstracts International $g 80-05B.
790 $a 0771
791 $a Ph.D.
792 $a 2018
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10930862