Beegoo.ru - new approaches to information search
Modern well-known search engines ( Google , Yandex , Yahoo , Rambler , etc.) use approximately the same approach to indexing information and organizing subsequent user search on it. The document encountered on the Internet is verbalized and the system associates the webpage with these words. When a user wants to find something, he enters a phrase and the search engine pulls out all the documents that he has associated with the words of this phrase. Then there are already nuances: to issue documents in the most relevant order, to take into account all forms of the word (plural, case, etc.), and so on.
There are at least two fairly common situations where such an approach does not allow a person to find what he wants.
Situation first. The user does not know what he is looking for. Suppose I need to find a birthday gift for my child friends. I know that this is a boy and that he is 3 years old. But I have no idea what children are playing at this age and what the modern children's goods industry offers. What should I do? Introduce a “toy for a boy of 3 years from 700 to 1000 rubles” to Google? Tried - does not help. A bunch of heterogeneous information indexed at different times. And I need actual for today, specific offers of specific stores. If you forget about the Internet for a moment and imagine what I would do in the real world, then most likely I would go to a large toy store and ask the seller the above question. And he had already digested my request, would lead me to a specific shelf of goods, reasonably limiting my search to suitable copies. I want something the same, but on the Internet ...
Situation two. Documents on the Internet do not always contain complete information. Returning to the subject of toys, we give another example. Suppose that the seller of the online store on his website wrote “Machine Lego Duplo”. Obviously, by the requests “Toy for a Boy”, “Designer”, etc. I will not find this document in the search engines that I mentioned above. On the other hand, in the real world, if I saw this toy in the catalog or “live” in the store, then it would be obvious to make obvious conclusions that it was for the boy and that Lego, the designer, would be difficult for me. And I would consider this proposal as one of the best.
What is it all about? The situations and problems described above are fairly common. And the more people will become Internet users, and the more they look for goods to buy there, the more acute these problems will be.
In our project beegoo.ru we tried to solve these problems using a certain approach to indexing information and organizing search. What is its essence? We have combined the capabilities of traditional indexing “by words” with knowledge about the subject area to which the information relates. The concepts of this subject area and various types of communication between them are fixed in this knowledge. And now the phrase from the document cited as an example above, we can relate not just to the words “machine”, “Lego”, “Duplo”, but add a lot more to this list:
')
Now, if the user contacts us with the request “designer for a child is 2.5 years old”, then we will be able to find this suitable product offer.
Such an approach to the organization of information also makes it possible to offer the user a mechanism for related prompts. Using it, the visitor of the search site will be able to construct a query in phrases understandable to him:
Our project is about six months. During this time we built a prototype and prepared for the service knowledge about children's toys. Now the service automatically scans online stores and parses their product offers according to the principle described above. Automatism reduces our costs for filling the information base and will give significant advantages when working with users - we will be able to provide them with the most complete database of offers on products of a certain subject.
In the near future we plan to add to beegoo.ru knowledge about children's clothing and shoes. After that, you should get an interesting resource for parents, which can take the place of a permanent tool for finding products for your baby.
In general, we invite everyone to us - www.beegoo.ru .
There are at least two fairly common situations where such an approach does not allow a person to find what he wants.
Situation first. The user does not know what he is looking for. Suppose I need to find a birthday gift for my child friends. I know that this is a boy and that he is 3 years old. But I have no idea what children are playing at this age and what the modern children's goods industry offers. What should I do? Introduce a “toy for a boy of 3 years from 700 to 1000 rubles” to Google? Tried - does not help. A bunch of heterogeneous information indexed at different times. And I need actual for today, specific offers of specific stores. If you forget about the Internet for a moment and imagine what I would do in the real world, then most likely I would go to a large toy store and ask the seller the above question. And he had already digested my request, would lead me to a specific shelf of goods, reasonably limiting my search to suitable copies. I want something the same, but on the Internet ...
Situation two. Documents on the Internet do not always contain complete information. Returning to the subject of toys, we give another example. Suppose that the seller of the online store on his website wrote “Machine Lego Duplo”. Obviously, by the requests “Toy for a Boy”, “Designer”, etc. I will not find this document in the search engines that I mentioned above. On the other hand, in the real world, if I saw this toy in the catalog or “live” in the store, then it would be obvious to make obvious conclusions that it was for the boy and that Lego, the designer, would be difficult for me. And I would consider this proposal as one of the best.
What is it all about? The situations and problems described above are fairly common. And the more people will become Internet users, and the more they look for goods to buy there, the more acute these problems will be.
In our project beegoo.ru we tried to solve these problems using a certain approach to indexing information and organizing search. What is its essence? We have combined the capabilities of traditional indexing “by words” with knowledge about the subject area to which the information relates. The concepts of this subject area and various types of communication between them are fixed in this knowledge. And now the phrase from the document cited as an example above, we can relate not just to the words “machine”, “Lego”, “Duplo”, but add a lot more to this list:
')
Now, if the user contacts us with the request “designer for a child is 2.5 years old”, then we will be able to find this suitable product offer.
Such an approach to the organization of information also makes it possible to offer the user a mechanism for related prompts. Using it, the visitor of the search site will be able to construct a query in phrases understandable to him:
Our project is about six months. During this time we built a prototype and prepared for the service knowledge about children's toys. Now the service automatically scans online stores and parses their product offers according to the principle described above. Automatism reduces our costs for filling the information base and will give significant advantages when working with users - we will be able to provide them with the most complete database of offers on products of a certain subject.
In the near future we plan to add to beegoo.ru knowledge about children's clothing and shoes. After that, you should get an interesting resource for parents, which can take the place of a permanent tool for finding products for your baby.
In general, we invite everyone to us - www.beegoo.ru .Source: https://habr.com/ru/post/116050/
All Articles