Feng Wu

Feng Wu

Auto-generated profile, based on publicly available information

Feng Wu

Xi'an Polytechnic University,School of Computer Science,Xi'an,China

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2021 in Expert Systems with Applications
Verified
Journal Article
Closed access
APAntonio Rafael Sabino Parmezan
Antonio Rafael Sabino Parmezan
HLHuei Diana Lee
Huei Diana Lee
NSNewton Spolaôr
Newton Spolaôr
FWFeng Chung Wu
Feng Chung Wu

7 August 2013 in Computer and Information Science
Verified
Journal Article
Open access
PDF file
YLYuanchao Liu
Yuanchao Liu
FWFeng Wu
Feng Wu
MLMing Liu
Ming Liu
BLBingquan Liu
Bingquan Liu
Presently, sentence-level researches are very significant in fields like natural language processing, information retrieval, machine translation etc. In this paper we present a practical task on sentence classification. The main purpose of this work is to classify the abstract sentences of scientific papers in the corpus built by ourselves into four categories- the background, the goal, the method and the result- which differ from each other in common usage, so that we can do further researches such as frequent pattern mining, information extraction and making a corpus for writing assistant system of scientific paper with these results. The main method of the classification is the Support Vector Machine, which is acknowledged among the best machine learning methods in the common text classification tasks. A semi-supervised method, Transductive Support Vector Machine, is also introduced into this four-class classification task to improve the accuracy. The experiments are conducted upon the corpus made by ourselves that consists of abstract sentences of scientific papers. The accuracy of the classifier finally reaches 75.86% with the semi-supervised method.

1 April 2026 in Journal of Health Informatics
Verified
Journal Article
Open access
AMAna Paula Merencia
Ana Paula Merencia
HLHuei Diana Lee
Huei Diana Lee
WTWeber Shoity Resende Takaki
Weber Shoity Resende Takaki
NSNewton Spolaôr
Newton Spolaôr
AFAlexandre Peiter Ferraz
Alexandre Peiter Ferraz
WCWu Feng Chung
Wu Feng Chung
Objetivo: Avaliar o desempenho dos métodos Symbolic Aggregate approXimation e Symbolic Fourier Approximation na representação simbólica de séries temporais de movimentos articulares humanos. Foram considerados três tipos de movimento (flexão/extensão, abdução/adução e circundução) e três tamanhos de alfabeto (três, sete e dez símbolos). Métodos: Oito voluntários saudáveis realizaram os movimentos com um smartphone fixado ao punho para coleta de aceleração. Os sinais foram suavizados com filtro Gaussiano (σ = 2), segmentados e convertidos em palavras simbólicas pelas abordagens avaliadas. Resultados: O desempenho dos métodos variou conforme o movimento. Para abdução/adução, Symbolic Fourier Approximation apresentou menores distâncias médias de Hamming, com diferença estatisticamente significativa. Para flexão/extensão, Symbolic Aggregate approXimation foi significativamente superior. Para circundução, Symbolic Fourier Approximation apresentou menor média, porém sem superioridade estatística. Na síntese entre as melhores configurações por movimento, a menor distância média global foi observada em uma configuração do Symbolic Fourier Approximation (circundução). Conclusão: A avaliação indica que não há um método universalmente superior; a escolha do método e da parametrização deve considerar o tipo de movimento e a consistência entre sujeitos. Esses achados apoiam a seleção de configurações específicas por movimento em aplicações de saúde digital, como monitoramento remoto, mHealth, Telessaúde e apoio à decisão clínica.

1 March 2022
Verified
Conference Paper
Closed access
SLShiyi Liu
Shiyi Liu
BYBo Yang
Bo Yang
FWFeng Wu
Feng Wu
CLChri Liu
Chri Liu
Inspired by the efficiency of dictionary learning and deep learning, we propose a novelty nonlinear feature representation method, namely Joint Deep Dictionary Learning (JDDL). JDDL embeds an ingenious design of network module which could achieve dictionary learning into common deep learning network structures. Specifically, JDDL learns the dictionary and sparse representation coefficients simultaneously in a low-dimensional latent space built upon deep auto-encoders. Hence, JDDL could produce compact and discriminative representation for the given data and even unseen data by incorporating dictionary learning and deep learning into a unified framework. Extensive experiments are conducted on four real-world data sets clustering to show that the proposed method could provide performance superior to many state of-the-art approaches.