Nabc Analysis of university application guide based on Bing search engine and Microsoft academic search

PreviousAs soon as we entered the scientific research circle, we ushered in the software engineering project, which lasted for nearly three months. The Team Project composed of five members for nearly two months is indeed not a trivial matter, it is very difficult to determine a project that is practical, innovative, and suitable for the team. At present, Microsoft academic search and Bing search engine resources are available, so I thought about the university application guide, and my ideas and objectives are not perfect. Now I am using the nabc method to analyze this, please give me more comments and comments.

1) n (demand for need: What needs does this idea solve for users ?)

Every autumn, especially when I was a senior, I began to prepare for various outlets. Most of them chose to continue learning. So it is a headache for everyone to select the master or Ph. D. Especially when we choose to go abroad and apply for any school or professor, we have to spend a lot of time exploring it. Which school is doing well in scientific research? Which one is better? What are the requirements of the School for student application? All kinds of information need to be summarized and organized. Taking studying abroad as an example: According to preliminary statistics by UNESCO, the number of international students worldwide in 3.43 million was 2.96 million, an increase of 12% from in the previous year. In China, there are now 0.3 million overseas students every year. So many people are not very familiar with going abroad, and each time they go to a search engine, the information is too noisy. In addition, some organizations that use commercial purposes will quietly enter the top of the search rankings, so it brings us a lot of confusion. What should I do?

 

2) A (approach practice: Does our team have a good solution? Is there a unique trick for our software innovation ?)

Users' pain points have been found. How can we filter massive amounts of information and obtain what we want? I suddenly realized that the final object we chose must be the teacher, so the teacher is the object we lock. What do we want from our teachers?

At present, my colleagues have also discussed many such things. When talking about teachers, we can basically see that teachers are not good, their scientific research level is not strong, and their scientific research is in a rising or stable period, or are you willing to give the students any sheep if you are powerless? Are teachers widely connected? How many projects does the instructor have? Is the teacher nice? Does the teacher take you when you don't understand it? Campus geographic environment? Consumption level? Lab information? Lab member? School application deadline? What is the school's online application process? What is the scholarship system for schools? Of course, there are still requirements for GRE and TOEFL.

The above is the idea of the masses, which can basically be divided into 1: the level of scientific research of teachers. 2: Relationship between instructors in scientific research.

3: individual students trained by instructors. 4: lab information. 5. school-related application regulations and procedures. 6: Scholarship system. 7: hardware requirements.

From a human perspective, this is a solution, but from a computer perspective?

Currently, machine learning is the most popular. Let's let him learn. You need to know that a person does not exist in an isolated society. A person cannot exist independently from the society. From the context of your circle, you can see the value of your life and your work situation, roles in society. Teachers need scientific research, and scientific research needs to be publishedArticleOf course, engineers will also publish many articles. Who are you working with in scientific research? There may be students and colleagues. Social graph technology is also quite mature. For example, the graph analyzed in arnetminer

People have a large scientific research network, and we have a certain degree of Cognition: co-authors colleagues can actually be similar to the teachers studied, he can be thought of as a duplicate (repetition) of the Teacher. There are also students in the graph. Okay. What can we do if we find the link graph?

If I used a web search to confirm that my teacher was working as faculty in a certain university, I also found that my co-authors colleague was also in this university, then I believe that my teacher is in this university. Of course, if I see this teacher on the school Webpage, I will confirm that the teacher is in this school and even know what department you are in. Therefore, colleagues in co-authors can better determine the school and institution of the teacher. (In many cases, teachers have their own personal homepages and put labs and so on their personal homepages. There are only a few documents that can be crawled by schools. So co-authors plays a role ).

The co-authors student is also a microcosm of the teacher. The teacher's homepage, or the lab homepage, will certainly have student information. We can contact the teacher's students directly to get a lot of unexpected information. Therefore, how to find student information and contact information is certainly critical. In co-authors, you can find the teacher-student relationship to determine the student's name. The lab home page also contains student information, and the instructor's personal home page also contains basic student information, we should get more comprehensive information from multiple directions. If we don't get the information, we can use the Bing search engine to help us, in fact, we certainly have a lot of prior knowledge, names, and labs should be able to get it. So you can find some information from the search engine and then filter it by the previous knowledge!

If the relationship graph is large and has many branches, it indicates that you have many scientific research contacts and many others. If co-author contains many other schools, therefore, you are closely connected with external schools and may have many communication projects.

Currently, graph analytics is similar. Now we still have the Microsoft academic search resource ). There are a lot of resources, a lot of data, the instructor's papers, homepage, and co-author. The instructor's paper can also be obtained from his personal homepage. If the instructor's personal homepage contains information such as publication/citation, it is estimated that we can crawl it. We will integrate the two into one (comparison or compromise ).

Our current resources include: Bing search engine ). Because the information in the search engine is too complex, we cannot believe it all, so we need to perform a certain screening, so how to filter?

Because all of our limited objects are applied in academic circles and colleges, there must be many rules in academic fields, such as the application process, because the application process is almost the same across the world, in other words, the steps in the application process of at least two schools will certainly be similar if they are sorted by TF (term frequency, term frequency. Therefore, when using Bing to crawl webpages, the obtained results are sorted. We can first obtain certain training data (such as typical application templates ), then, sort again! Obtain the best information.

For GRE/TOEFL, I believe all major commercial websites will be available, so since the resources are there, we should be able to search for them.

Well, I have analyzed so much to see if our problem can be solved from the computer perspective? I proposed a rough solution at seven:

1: Instructor's Scientific Research Level: Instructor's personal homepage, Microsoft academic search

2: The teacher's scientific research cooperation relationship uses co-author to determine the teacher's institution, the teacher's scientific research contacts, and communication with other universities

3: the instructor's personal Student Training Status: Ask the students through the relationship graph or the student information in the lab or the instructor's personal homepage.

4: lab information: the lab homepage, which should be mentioned on the instructor's personal homepage.

5. Related school application regulations and procedures: Use Bing search and training data to filter applications.

6: Scholarship system: search by Bing, crawl the school website, and filter by training data.

7: hardware requirements: as many organizations and many webpages have such information, it is easier to obtain it!

At present, the teacher's problem has been solved. Now let's figure out our general process.

3) B (Benefit benefits: what benefits does this product/service bring to customers/users ?)

Of course, if you do well, you can at least serve as a reference for students. Our software is free, and we can also use this website to create academic forums to solve the most painful problems for students. We provide students with the most authentic information and the data they want as much as possible. In addition, we can collect statistics on the information and analyze the trend to further help students.

4) C (competitors competition: our competitors ?)

At present, many commercial organizations on the market are created to study abroad, and they are very successful. However, their biggest weakness is that they are not academic and profitable. We use a variety of tips, smart, it takes a lot of effort to make a lottery in the application process. They are sending you out now. This will often have a great impact on students. We collect all kinds of data from the original perspective of users for user query and user selection. Therefore, if you do well, it will certainly be very popular!

I have said so much. In fact, I would like to say that I have provided a sincere attitude to share my thoughts with you. I sincerely hope that you can provide valuable comments and criticize me, don't hesitate, welcome!