Exploring the path of social search-Xunlei

Exploring the path of social search-Xunlei

Debut: Social search weekly Author: Ma Guoliang

Ailog said that thunder is a successful social and social search model. However, every time I talked to my friends about the advantages of the thunder model, I had never thought about how to connect with social and social search; on the other hand, we have been recording and describing social and social search cases and models, but have not looked back. In fact, Thunder is also a special case of social and social search. Now, we can use thunder as the most successful examples of social and social search models.
I. Collections of Xunlei social search
The rapid growth of thunder is the first benefit from the excellent thunder download client. The second is because of Thunder's own resource database. Thunder's own resource database is actually a result of socialization. The use of the Xunlei client is similar to the tool bar launched by social search engine websites, but it is more advantageous than the tool bar, because the tool bar threshold is too low, the Xunlei client has always maintained a leading edge in similar products. The tool bar of the Social search engine website is used to submit favorites to the search engine. Users generally take the initiative to submit their favorite information; the role of the thunder client is that users submit collections to the thunder search or resource database, but users do not even know how to submit their favorite information. This is the brilliant place of the thunder product. Currently, when users use the thunder client to download resources, they also upload a copy of the download information to the thunder server, this ensures that the search results obtained through Thunder are rich and are the latest and most popular resource. With the increase of the client users, the thunder server can easily grasp the Internet user resource source download index.
Ii. resource library of Xunlei social search
Xunlei's resource library is rich and rich. As long as users continuously use the Xunlei client, the resource library will expand. The foundation of the thunder resource library is the targeted crawling of all major resource websites. Similar to social search engines, many of them are based on Meta Search and other search engines (domestic Meta Search socialization case: deyeb ). Therefore, Thunder's resource library consists of two parts: one is automatically captured by machines, and the other is generated by users. When a user generates a small amount of resources, machine search takes the lead; when a user generates a large number of resources, although the user may use the resources to meet user needs, however, the machine search crawlers still need to be supplemented as necessary to help search results. But what if there are too many invalid or junk Resources in the user-generated resources? How can we ensure that users can find what they need most by using search engines? This is a problem involving resource filtering and filtering.
Iii. Resource screening for Xunlei social search
While enriching the resource database, we must ensure that the resource information is excellent and benign. The social search engine is also facing the challenge of too much junk information submitted by users. If you are not careful, the social search engine becomes the "garbage bin search engine ". The traditional favorites or bookmarkdonefile websites inevitably interfere with junk information, and Yahoo's favorites cannot escape and become a "garbage bin. The 365key in China is also. I have been thinking about how the user screening mechanism works? As mentioned in my previous articles, favorites + SEARCH + dig are a perfect combination mechanism. Adding junk information to favorites cannot be excluded. Searching determines users' interests and needs, so that some junk information can be avoided. Dig can filter information and remove bad information from users' search results. After filtering by DIG +, search results irrelevant to users are greatly reduced. Once a benign user atmosphere is established, the search will also become a wiki search, and the search results will become very accurate. What kind of search is your favorite? How does Thunder filter resources? Users like a search engine, because the search can find the best information in the shortest time. The search mechanism of this search engine should include machine filtering and manual participation. Simply put, the priority of the collection site is set during machine collection. This is determined by factors such as the resource heat, network download speed, and so on. The resource heat is determined based on the user's attention and downloads. The Dig method used by the user for resources of thunder is to download. Each download is equivalent to a vote for the resource.
Iv. Resource clustering for Xunlei social search
In fact, as a dedicated resource search and download site, Xunlei seems to have little demand for the Social search function. You can do one related search at most. The resources and scope of thunder search are still limited. For example, when we use thunder to search for "Sister Furong", social search engines often use words such as "blog", "Photos", and "Peking University, these words can help users fully understand the comprehensive information of Sister Furong. The videos or images searched by thunder are nothing more than videos or pictures of Sister Furong, Which is what users need most, compared with social search, quick search has a narrow range. A specific purpose determines the difference between quick search and Social search. Thunder does not need to be clustered, but it is very important for social search. Classification is even a prototype of Smart Search in the future. Simple intelligent search involves heuristic methods, the search engine is used to help users find the best answer, and continuous guidance and prompts play a key role in the search process. The quality of classification determines the search accuracy and the amount of time consumed.
IV. Other features of social search
1. Personal Search
Why do people need to search for social search? Social search can be based on personal search. In this mode, swicki and rollyo are relatively successful in foreign countries. Use personal search as the breakthrough point, customize the search page and search results, you can also specify a search site, and then put the code on your personal website or blog. Because the user groups of personal websites or blogs are fixed, at the same time, custom search pages, search results, and specified search sites can meet the requirements of individual webmasters and blogs to customize a search engine for their own websites or blogs, the suggestions for adding, deleting, and modifying the user's search will also help improve personalized search for individual users, so personal search will be created one by one to meet the needs of the search. The recommendation of personal search will directly affect the search results of the overall search engine. Unfortunately, swicki has not yet launched a public search service. Swicki is indeed a wiki-based search. When we mention this feature, we will look at thunder again. The Socialization Mechanism of thunder is just getting started, but it has taken shape. We can see that the resource blog and webxunlei homepage are independent of each other. Webxunlei is similar to swicki's personal search, but this "Personal Search" is only for individuals. Downloading resources through webthunder is an update process for the Thunder resource library. Without your knowledge, you can edit the Wiki for the Thunder resource library. Because downloading resources may involve privacy issues, otherwise they can be taken out separately. each user has a resource library and a search resource to share his/her personal download resources. The resource library compiled by individuals is often very valuable and convenient for other users. When the active participation of domestic users is not very high, webxunlei will list a resource library for this user for each registered user, you only need to sort it out. It is meaningless to simply use a resource blog. A blog combined with webxunlei's user resource library may be more valuable. Isn't it a swicki personal search?
2. Circle
When talking about sharing, we certainly need the concept of circle. During the development of social search, we gradually summarize the shortcomings of common search engines and gradually improve them. The sharing of search is very necessary. The previous one searches for the same content, which provides guidance and help for the latter. There are many users who search for the same thing on the search engine every day. We call these users who search for the same thing as "interest circle users", and these user groups form the same interest circle. The first layer of the interest circle is to list keywords related to searches by other users, including related searches and similar keywords. However, this is a search hotspot generated by users. The second layer of the interest circle refers to users with common interests and share their search results. For example, Li Yuchun's corn must have a common search demand and search hobby. Communication between them is also necessary because they share a common interest point. The social search circle service is a deep service for users. It is the purpose of social search to better meet user needs. It is not enough for thunder to complete the circle in the resource blog. Blogs can be shared, but users cannot further communicate with each other. In addition, traces of search records can also be shared in circles. Of course, keywords can be blocked to protect your privacy. The ultimate goal of the circle is to form a cultural atmosphere for sharing and achieve interaction.

The last point is that this article only discusses the pattern of Xunlei and social search. Thunder is not a social search, but a social search is not a thunder. Because there are similarities, we can find some excellent applications from the Social search model, social search can also be used for reference in the rapid development of thunder.