How many times have you not found the right words to formulate your search query? How many times did the responses not meet your expectations? Search Insight – a semantic search engine developed by Makolab that significantly improves the quality and relevance of search results.
Semantic search has become a core focus in the field of artificial intelligence. This innovative solution for searching and analysing data sets provides a practical tool for receiving the most relevant information from search queries. The Search Insight solution developed by MakoLab is a module that has encompassed the ideas of semantic search completely, resulting in its effective use in a variety of commercial projects. One such project is the RSI-CMS, implemented by MakoLab on behalf of Renault-Nissan and successfully implemented in more than 60 markets around the world. The essential nature of the search engine feature in this CMS has meant that the development of a semantic search engine has in turn provided the capacity for higher quality results. Our first implementations of this semantic technology are looking very hopeful!
Solving challenges with semantic search
The standard search engine used on most websites analyses inputted phrases as a string of data entered into the search box and then returns search results that represent pages containing phrases that match the original query. The problem with this style of searching is that the user is likely to not know the structure of the webpage they are attempting to locate, moreover they do not know whether the information they require actually appears on the pages shown in the search results. For example: A user enters a search query for “security”, intending to find information concerning systems that will help ensure their safety while driving. By entering this, the user might expect the search engine to reveal pages containing information about such security systems. However, it may turn out that the information the user requires is actually found on sub-pages that do not contain the phrase “security”. In this case the user will not receive the relevant search results, although in the structure of the site itself there will be sub-pages with the information regarding security systems such as ABS, emergency brake assistance, ESC, etc.
Searching for “security” with Search Insight
Utilising a semantic approach instead provides the capacity to “understand” the intent of the user looking for information on a given page. Search Insight “analyses” what information the user is looking for and presents relevant results to them in a clear way. In this way, the user searching for “security” is able to locate the relevant sub-page in their search results, due to the attributes of the data on these pages being semantically related to the term ‘security’.
For Renault sites based on the RSI-CMS, an ontology dedicated to the automotive industry was prepared by Makolab. This ontology is a formal dictionary of terms related to the automotive industry- consisting of a collection of concepts and the relationships between them. As a whole, this ontology creates a conceptual scheme that constitutes a tool for the general description of the automotive industry.
A major advantage of the semantic search engine implemented on Renault’s pages is that it assigns search results to thematic groups. From this approach, the user receives information in a logical and transparent manner that can be divided by general information, technical data, customer service issues, sales offers et cetera.
The Search Insight administration module allows users to promote pages in the search results list and create synonyms for words ontologies – such as “Linda” for “Talisman” (referring to the fact that the actor Bogusław Linda was featured in the advertising campaigns of this car model). This means that by entering in the search engine window the phrase “Linda”, search results will return pages regarding the Renault Talisman car. As with the above example, ‘ABS’ can thus be categorised under “security”, benefitting the end user and the company in question tremendously by providing the user with the sought after page in their search results.
The first Search Insight implementations took place for Renault Malta and Renault South Africa in April 2017.