MatrixSearch Engine DECODED

MatrixSearch Engine provides powerful search options against any string, which makes it a precise medium of searching nearest strings. This recommendation engine brings together the power of search and spellcheck; engine searches nearest word match in any unstructured string.

MatrixSearch corrects spelling mistakes with unmatched accuracy. Spelling checker uses a Levenshtein Distance Algorithm which is based on matrix technique to correct any type of mistake and set priorities as per matching strings.

Example

  1. HONDA PORTUGAL SA

M1

  1. M 6 COUPE

M2

  1. HONDA

M3

  1. Media General Promotion

M4

Application

1. Provide auto string mapping against source string.
2. Provide matching recommendation against source string.

Business Needs

Search engine increases your chances of discovery when customers search for a product or service with Google, Bing, and other search engines.

For any custom ETL tool, focus is to get unmapped entities from staging area and show best matching strings as per previous saved translations. We are writing custom ETL service to handle complex business logic with multiple file templates and languages as per data provider companies.

We developed our own custom recommendation engine to achieve complex file processing using Levenshtein Distance Algorithm.

Note: MatrixSearch engine works against text only. (At the time of writing, it does not support any media files, for e.g. picture, html and video).

Transform Techniques

Some of the most powerful approaches to semantic processing are based on the use of mathematical transform techniques. Matrix driven techniques have been the most successful. Levenshtein Distance Algorithm, based on matrix technique is used to obtain high precision results.

MatrixSearch Processing

MatrixSearch engine as per following cases:

  1. Word to word.
  2. Word to string / string to word.
  3. String to string.

Text Process comprises of following steps:

  1. Receive source text key.
  2. Target text acquisition.
  3. Compare, Refine, Evaluate and Iterate.
    Compare:
    1. Compare source key text with target text.
    Refine:
    1. Define content strategy.
    2. Sort as per nearest match rank (measures effectiveness).
    Evaluate:
    1. Define content size for result set.
    2. Content creation.
  4. Pass information for business use.

MatrixSearch Engine

Summary

This was an overview of the MatrixSearch engine. A Search/Recommendation engine can be a complex topic. We have implemented small and important parts of the search engine as per our business needs. Your business needs to have a search strategy in place if you are providing/harmonizing online marketing data. You are welcome to share your views as comments to this blog.