Selected Projects

Weenix Operating System

Implemented kernel processes & thread, virtual file system, virtual memory of an operating system including but not limited to mutex lock, context switch and wait, path resolution, address spaces. Passed over 1000 tests on both kernel-level and user-level thread. code

Distributed Lookup System

In this project, my contribution is consist of two parts

Implemented a distributed lookup system from scratch. It contains three backend databases, an intermediate server, a client and a monitor and established stream sockets to transmit queries and results among all devices.
Implemented the methods of exact matching, prefix matching and suffix matching for backend server. code

Smart Home System using Rasperry Pi

In this project, we design prototype of a smart home system based on Raspburry Pi.

We make full use of sensors to measure multiple environmental parameters
Based on these parameters we can turn on or turn off the electrical devices like light or alarming system automatically.
we combine the data-mining techniques into the prediction part of our system, making our system more intelligent.

More details can be seen here

Link Prediction using Textual and Graphical Features

In this project,we mainly focus on link prediction on the scholarly data. The dataset contains citation networks of 27770 documents with the information of title, abstract, authors and publish time. It is the same data using in AXA Data Science Winter School : Tsinghua, Renmin and Ecole Polytechnique

In this data challenge, I utilized 19 features and build a boosting system to predict the citing probability. The 19 features are:

Topic representation selected by PLSA
Lsa representation of abstract
Lsa representation of topic
Tf-idf representation of abstract
Tf-idf representation of title
Tw-idf representation of title
Tw-idf representation of abstract
Number of common words in abstract
Number of common words in title
Community of papers
Pagerank score of cited paper
Pagerank score of citing paper
Clustering score of cited paper
Clustering score of citing paper
Degree of cited paper
Degree of citing paper
Number of common author
Time gap I get the f1 score of 0.9765