| Have you ever wondered how Search Engines (SE) | | | | mathematical form that the computer can understand |
| know which pages to give you when you do a | | | | and process. |
| search? For example when you search for "Apple" | | | | The use of semantics is really central in LSI for SE. |
| the computer, how does the SE know not to give | | | | They are used to try to emulate what a human |
| you pages about Apple juice or Apple farming? | | | | being would rank highly. Semantics refers to the |
| The solution that handles this is what is often | | | | aspect of meaning in a language. Its here we get |
| referred to here as a SE algorithm. An algorithm is | | | | terms like synonyms and polysems (i.e. the same |
| mathematical formula for solving complex problems. | | | | word with different meanings e.g. Apple for the |
| SE algo is the formula used by SEs to rank the pages | | | | Macintosh computers and Apple for the fruit) |
| contained in its index depending on the query. Years | | | | The question doing rounds in SE forums around the |
| back when SE algos were still rudimentary they used, | | | | internet is whether or not Google is using LSI in their |
| what in retrospect are very crude means to provide | | | | ranking. What is known for sure is that Google Inc. |
| most relevant results to a query. Among them was | | | | did buy Applied Semantics, a company that creates |
| the use of keyword density. This used frequency of | | | | software applications based on LSI technology. We |
| a particular word or term to decide if it should rank | | | | also know that LSI is in their systems. How do we |
| higher or lower. | | | | know you ask? Just use the toggle key before your |
| But unfortunately this does not represent complete | | | | search query. The toggle key is the key often to the |
| reality. For example it was the assumption of the | | | | extreme left of the computer keyboard just below |
| algo that if a page is about dieting, the word should | | | | the escape key with the "~" sign. You will need to |
| be the most frequently used word apart from other | | | | hold down the shift key. |
| grammatical necessities like me, I, and, the, or etc. | | | | If you type it before the term "apple" you will notice |
| But this is not always true. This is because in most | | | | that in the resulting SERP are about computers with |
| writings, to avoid monotony there is a mix of words | | | | words like Mac, Macintosh, G4, Apple Computer |
| of similar meaning - synonyms. So a page about | | | | bolded. If you toggle the plural i.e. "apples" you will |
| dieting may also include lose weight or weight loss. | | | | notice that now you have pages on apple fruits and |
| Enter LSI; LSI or Latest Semantic Indexing is an | | | | the word fruit is also bolded. The bolded words are |
| informational retrieval system that depends on a | | | | what the SE also considers as synonyms to the |
| technique used to process natural language into a | | | | search queries "apple" and apples" respectively. |