Motive imagery scoring based on deep neural network

The purpose of this study is to develop a method for scoring the motive imageries in text materials. According to Winter’s motive scoring system, there are three different kinds of motive imageries and each of them is given detailed definitions and scoring rules. But it’s difficult and also time-con...

Full description

Bibliographic Details
Main Author: Yu, Sicheng
Other Authors: Chen Lihui
Format: Thesis
Language:English
Published: 2019
Subjects:
Online Access:http://hdl.handle.net/10356/78596
_version_ 1826120408524390400
author Yu, Sicheng
author2 Chen Lihui
author_facet Chen Lihui
Yu, Sicheng
author_sort Yu, Sicheng
collection NTU
description The purpose of this study is to develop a method for scoring the motive imageries in text materials. According to Winter’s motive scoring system, there are three different kinds of motive imageries and each of them is given detailed definitions and scoring rules. But it’s difficult and also time-consuming to implement these rules manually. The traditional machine learning methods also have difficulties in extracting features. With the evolution and development of deep learning, deep neural networks have played an important role in data processing. In the study, three different deep learning models, including TextCNN, LSTM and Bidirectional LSTM with attention mechanism, are applied to score the motives. The performances of three models are evaluated, compared, and reported in this thesis.
first_indexed 2024-10-01T05:15:58Z
format Thesis
id ntu-10356/78596
institution Nanyang Technological University
language English
last_indexed 2024-10-01T05:15:58Z
publishDate 2019
record_format dspace
spelling ntu-10356/785962023-07-04T16:19:20Z Motive imagery scoring based on deep neural network Yu, Sicheng Chen Lihui School of Electrical and Electronic Engineering DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing The purpose of this study is to develop a method for scoring the motive imageries in text materials. According to Winter’s motive scoring system, there are three different kinds of motive imageries and each of them is given detailed definitions and scoring rules. But it’s difficult and also time-consuming to implement these rules manually. The traditional machine learning methods also have difficulties in extracting features. With the evolution and development of deep learning, deep neural networks have played an important role in data processing. In the study, three different deep learning models, including TextCNN, LSTM and Bidirectional LSTM with attention mechanism, are applied to score the motives. The performances of three models are evaluated, compared, and reported in this thesis. Master of Science (Signal Processing) 2019-06-24T05:38:59Z 2019-06-24T05:38:59Z 2019 Thesis http://hdl.handle.net/10356/78596 en 69 p. application/pdf
spellingShingle DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
Yu, Sicheng
Motive imagery scoring based on deep neural network
title Motive imagery scoring based on deep neural network
title_full Motive imagery scoring based on deep neural network
title_fullStr Motive imagery scoring based on deep neural network
title_full_unstemmed Motive imagery scoring based on deep neural network
title_short Motive imagery scoring based on deep neural network
title_sort motive imagery scoring based on deep neural network
topic DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
url http://hdl.handle.net/10356/78596
work_keys_str_mv AT yusicheng motiveimageryscoringbasedondeepneuralnetwork