What is the relationship between search and referral systems in a Web site?

When we open any portal site, we will find that there are site searches in the website, that is, the site search, its role is to help users quickly find the information we want, however, when we open a message on this page we will also see the relevant recommendations for our users to choose. However, this site search and information recommended the role is to better improve the site's user experience, do not know if you have any idea whether there is any correlation between the two?

This feature was found early in the search, I believe we should all be very familiar with, then we will focus on the following introduction of this later recommended function, we must fully understand the characteristics of the recommendation to better see the two directly what the advantages and disadvantages have been directly related to each other.

What is a referral system?

The historical background of the search engine or retrieval system is basically how to find the part of the user in the face of information overload. For example, what should we do if we need to purchase a new server to meet the needs of the use? Suppose there is an HP agent nearby that can be called directly or in his shop to find the right model. Or if we don't want to go out, we can also open the HP or Dell website, find the server in their Product catalog list, or find the server directly in the search box. In these two examples, procurement in fact, there are some information overload, but the store may be very small, find the appropriate models do not need additional help, on the site because there are many different models to choose from, visitors need to rely on classification information or search bar to find their own models. The core of the two examples here is not the degree of overload, but the key node is that the user or the visitor has a clear need of their own, at least the caller can describe the information autonomously.

But what if a visitor or user doesn't have a clear need? For example, when I write this text, I want to find some background music, but if I directly open a music site, such as Baidu's MP3 channel, face thousands of albums, it is difficult to say I still have a clear choice of direction, often overwhelmed, Don't know where to start. At this point, I am faced with an information overload problem, is essentially a very serious and widespread information overload, but search engine or retrieval system has been very difficult to give me direct and effective help. At this point, I need an automated tool that starts with my music history data and finds songs from a huge library that match my listening habits and preferences to my choice, a tool that can be called a recommendation system.

The essential difference from the search engine

In order to solve the problem of information overload, countless tools and solutions have been invented, so far, to do a better job, really get most of the user identification methods can be divided into two categories: Classification directory and search engine. Classification directory by artificial or half artificial way on the Internet site classification to facilitate the search for information about certain industries, but obviously with the speed of the human brain is far from being able to keep pace with the development of the Internet. Search engines let users find their own needs in the form of keyword information, is to solve the problem can not match the whole network, but still faced with the need for users to start the problem, that is, if the user can not accurately describe their needs, the search engine can not work or can only work in a very inefficient way.

The recommendation system is consistent with the classification directory and search engine objectives, he is also a tool to help users find useful information more quickly, but unlike the search engine, the recommendation system does not require the user to provide clear requirements, the recommendation system will itself from the user's historical behavior data, for the user's needs and interests to build models, It is based on this to match the user's needs with a larger amount of information. Therefore, from this point of view, search engine and recommendation system is essentially complementary two tools, search engine to meet the needs of users with a clear need for active search needs, the recommendation system to meet the user in the absence of clear needs when the information to explore needs.

How does the recommendation system work?

In order to better understand the difference between the two, it is necessary for us to make a simple introduction to the basic principle of the recommendation system, we review the real life of how we face a variety of choices to make decisions (information overload in the case of the decision), to explain the basic principle of the recommendation system. The following is still an example of listening to music, in general we may use one or more of the following methods to solve this problem:

Consult our friends and colleagues who we think are more professional. We might have a chat with a friend who often listens to music, ask him if he has any great albums available recently, ask them if they have anything to recommend, and if we have enough fans and are active enough, we can also send a "please recommend music to me" on our microblog., and then wait for the fans to help. This is probably our most commonly used recommended method, referred to in the recommendation system as a social recommendation, that is, let friends recommend things to themselves.

However, through the above recommended methods, it is not difficult to see the nature of the recommendation system through a certain way in the goods, content and users to establish links between the nature of the search engine and the task is very different. This article is edited by http://www.gkdai.com, reprint please specify!