USTC

 

Computer Science
                                                  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wenhua Cheng(程文华)

 

Master of Computer Science

 

 

Add.:    Depart. of Computer, USTC , Hefei, Anhui, 230027, P.R.China 

Email:   cwh AT mail DOT ustc DOT edu DOT cn

Phone:  +(86)18725515363

 

 

Research Interests

-         Data mining  

-         Bioinformatics

-         Parallel Programming Model

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Education

 

I obtained B.S. in Department of Computer Science and Technology in Northeast Normal University(NENU) in 2010. Next year, I was recommended for admission to the University of Science and Technology of China(USTC) without joining Graduate Candidate Test. My supervisor is Associate Prof. Yun Xu. In the first-grade of post-graduate, I have learned many advanced courses of Computer Science, for example Parallel Programming, Parallel Algorithms, Probabilistic Algorithm and Distributed Algorithm, Advanced Operation System, Advanced Software Engineering, Combinatorial Mathematics, etc. In the same year, I participated in data mining competition, and won the first prize. In the second-grade, My research interest is in Bioinformatics.

 

 

 

 

 

 

 

Research Projects

 

 

 

A New Accurate Marriage-Friends Recommendation System

 

Wenhua Cheng, Pengyu Nie      First Prize of China

 

In today's society, many young people busied themselves at engaging in social work and learning. They have no time to know their boy or girl friends. They always make friends in marriage dating site, for example Jiayuan. A good marriage-friends recommendation system can improve the accuracy of recommending friends who they want to know. In a marriage dating site, many person can be recommended to one person, the person can select to give a click or not, give a message or not. Now, we have been given many information of recommendation and information of clicking or giving message.  We must predict the persons in which they may have interests. In this competition, we designed a mathematical model, based on the information of clicking or giving messages. If a person was clicked or sent messages frequently, he or she has a higher charm value. Adding to the information of age, the model which we designed gives the accuracy result of recommendation.  In the end, our NDCG@10 is 2.53, outperforms which all other terms get.

 

 

 WinHAP: an Efficient Haplotype Phasing Algorithm based on Scalable Sliding Windows

 

Yunxu*, Wenhua Cheng, and Pengyu Nie.

 

WinHAP is a fast phasing program, based on scalable sliding windows and maximum parsimony principle. It allows a rapid and more accurate inference of haplotype from population genotype data. We have known 2SNP(Bioinfromatics, 2006) software package implements a very fast scalable algorithm for haplotype inference based on genotype statistics collected only for pairs of SNPs. WinHAP is the simple generalization of the 2SNP which just using pairs of SNPs. It uses more SNPs to inference haplotypes and the speed is almost the same. Empirical evaluation shows that WinHAP outperforms most of the existing algorithms in terms of both speed and accuracy for large data sets with thousands of individuals.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Awards

 

1.      First Prize of the 1st Collegiate Data-mining Competition of China in 2011 (rank 1)

2.      First Prize of Northeast China Multiprovincial collegiate Progamming Contest (ACM/ICPC) in 2008  (rank 6)

3.      First Prize of Jilin Province Collegiate Progamming Contest (ACM/ICPC) in 2008 (rank 3)

4.      Second Prize of Asia Loval-National Contests Northeast China (ACM/ICPC) in 2009 (rank 15)

5.      Second Prize of Jilin Province Collegiate Programming Contest (ACM/ICPC) in 2007

6.      Second prize of Jilin Province Collegiate Programming Contest (ACM/ICPC) in 2009

7.      President Scholarship of NENU in 2008

8.      President Scholarship of NENU in 2009

9.      First Prize Scholarship of NENU in 2007

10.  First Prize Scholarship of NENU in 2010

 

 

 

 

 

 

 

 

Software

 

WinHAP:  It is a fast and accurate phasing program, based on scalable sliding windows and maximum parsimony principle. It outperforms most of the existing algorithms in terms of both speed and accuracy for large-scale data sets with thousands of individuals.

 

 

 

 

 

 

 

Assisting Course

 

1.      Design and Analysis of Algorithms (for engineering postgraduate in CS and SSE): Fall 2011

2.      parallel computing (for undergraduate in CS): Spring 2012

 

 

 

 

 

 

 

 

Publications

 

1. Yun Xu, Wenhua Cheng, Pengyu Nie, Fengfeng Zhou. (2012) WinHAP: an Efficient Haplotype Phasing Algorithm based on Scalable Sliding Windows.   [PDF]