Post by amirmukaddas on Mar 12, 2024 14:34:45 GMT 5
I'm doing a study on what I call "semantic vectors." It's still early to describe the results, in the meantime I'll tell you about the theory and practice of an experiment. semantic vectors Every time I talk about semantics applied to search engines , an SEO with at least 10 years of experience is struck by widespread unease, even if I do nothing but reflect on the mechanisms through which Google establishes relationships between terms present in one or more web documents and the documents themselves. The aim of my reflections is to understand if and to what extent these relationships help Google to recognize precise meanings, such as to improve the understanding of web documents and weigh in the attribution of rankings in search results.
In short, if Google attributes meaning to a term based on other terms present in the same context (article), in the same way it could attribute meaning and value to an entire article based on related terms present in other articles . It's not immediate, I'll try to explain myself better. Vector geometries Imagine that in document A the term table is placed alongside other terms, such as to design a particular wordspace . The terms Denmark Telegram Number Data have to do with prices , types , materials , manufacturers , styles , combinations , rooms . Google will probably understand that we are talking about a guide to choosing and purchasing a table and that we could be in the sphere of furniture e-commerce. There isn't much else to understand, also because we need to consider that Google simply associates terms without interpreting them .
Now move your attention to another context in which we talk about furnishing the house, but using different wordspaces, the wedding context . Very often those who organize their wedding are purchasing furniture for the house where they are going to live in the same period. He imagines that documents B, C, D... N exist in the various portals on marriage, developed precisely on this topic. What would happen if Google found a two-way match between terms used in document A and those present in other documents related to the wedding? In practice, by dividing document A "table purchasing guide" onto documents B, C, D... N "choosing furniture according to the organization of the wedding", you can bring document A closer to the others, making it more relevant compared to the query “wedding”, more so than other documents on the internet that explain how to choose the table.
In short, if Google attributes meaning to a term based on other terms present in the same context (article), in the same way it could attribute meaning and value to an entire article based on related terms present in other articles . It's not immediate, I'll try to explain myself better. Vector geometries Imagine that in document A the term table is placed alongside other terms, such as to design a particular wordspace . The terms Denmark Telegram Number Data have to do with prices , types , materials , manufacturers , styles , combinations , rooms . Google will probably understand that we are talking about a guide to choosing and purchasing a table and that we could be in the sphere of furniture e-commerce. There isn't much else to understand, also because we need to consider that Google simply associates terms without interpreting them .
Now move your attention to another context in which we talk about furnishing the house, but using different wordspaces, the wedding context . Very often those who organize their wedding are purchasing furniture for the house where they are going to live in the same period. He imagines that documents B, C, D... N exist in the various portals on marriage, developed precisely on this topic. What would happen if Google found a two-way match between terms used in document A and those present in other documents related to the wedding? In practice, by dividing document A "table purchasing guide" onto documents B, C, D... N "choosing furniture according to the organization of the wedding", you can bring document A closer to the others, making it more relevant compared to the query “wedding”, more so than other documents on the internet that explain how to choose the table.