发现了一个挖掘科研网的工具:

是由清华大学计算机系工程研究室开发的，感觉不错

Knowledge Discovering (19/30)	The main goal is to search expertise. (When you type a keyword, the engine output the experts for that keyword). Search result contains expertise's profile. List a lot of topics, show the expertise and events of the topic. Support ranking the authors by different attribute. Support ranking the conference according to impact factor. And showing conference trends. Support paper search. (When you type a keyword, the engine output the paper in right side.
Reference Management (12/15)	Contains reference for some paper. Contains citation for some paper. Contains different format to cite the paper.
Personalization (0/10)	ArnetMiner provides you a personal accounts and personal page. (few people use it).
User Experience (13/25)	The interface of this web page is complex. Too many words and content in each page. The responds time for each query is not very fast. (larger than 1 second).
Socialization (18/20)	Social graph: three relationships between each pair of authors: advisor, advisee, and coauthor. Contains course search, show course and the teacher. (However, some links are out of date) Addition: Social graph can find paths between two different teachers. The interface is good.
Users	For students to find their favorite professors. For people evaluate professors according to different rankings.

Total:62.

 

Details:

As the introduction in ArnetMiner said "ArnetMiner aims to provide comprehensive search and mining services for researcher social networks". It is different from many other academic search such as: Google scholar.

 

Knowledge Discovering: The main function aims to find a researcher. For example, type Bruce Croft and press Enter. The page show Bruce Croft's profile. The profile contains: positions, affiliation, address, phone, email, homepage, research interest and many othings.

For each researcher, you can find his friends, students or co-authors in right side and find the conferences which he published a lot of papers.

In order to evaluate the professors, you can rank the professors according to different attributes. There are 8 different ranking attributes: H-index, Citation, Uptrend, Activity, Longevity, Diversity, Sociability, and New Star.

For example, we sort the expertise in machine learning area according to H-index:

It shows a list of researchers. Unfortunately, there are some researchers' picture unfound. To find more interest things, you can go through the link: 

 

Additionally, it can rank the conference and analysis the authors' nationality distribution. You can go through the link: 

 

Reference Management: For each paper, you can find the abstract, reference and cited paper in the ArnetMiner. When you want to cite this paper, it provides a lot of format for you. For example, for paper "Click Chain Model in Web Search", you can cite it by following ways:

This function makes it easy to manage the references of our paper.

 

Personalization: You can login and create your page. However, I find there is only a few people use this function.

 

User Experience: The interface of the main function is complex, which contains too many words in one page.

 

Socialization: As I mentioned above, this engine mainly focus on search research social network. For each researcher, there is a social graph:

You can filter unimportant relationships by the number of co-authored papers. The relationships in different colors have different meanings. However, there are some error relationships.

Recently, there is a new function that is to find the relationship chains between two researchers. It can help us finding a strong recommendation letter.

 

Users: I think people who use ArnetMiner aim to evaluate the professor according to different evaluation metrics.

 

Author: Dong Wang