Native Apps At The Client & Cloud

Srinivasan Sundara Rajan

Subscribe to Srinivasan Sundara Rajan: eMailAlertsEmail Alerts
Get Srinivasan Sundara Rajan: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: MySQL Journal, Web Analytics, Cloud Data Analytics, Microsoft Developer, Big Data on Ulitzer

Article

Sentiment Analytics in SQL Server

Semantic search builds upon the existing full-text search feature in SQL Server,

Sentiment analytics is an emerging technology enabler for enterprises to extract the sentiments, opinions and emotions from their Big Data sources, so that enterprises can predict the potential acceptance of their products and offerings. This goes a long way in defining the success of the enterprise. As mentioned in my earlier posts it's always better to explore the extensions from the existing investments so that the enterprises get the best out of their existing investments. In this context the following notes throw some light on how sentiment analytics can be performed with SQL Server databases.

Semantic Search in SQL Server
Semantic search builds upon the existing full-text search feature in SQL Server, but enables new scenarios that extend beyond keyword searches. While full-text search lets you query the words in a document, semantic search lets you query the meaning of the document. Solutions that are now possible include automatic tag extraction, related content discovery, and hierarchical navigation across similar content.

There are two major functionalities that are part of Semantic Search Features, which can help when doing Sentiment Analytics.

  • Finding the key phrases in a document, ‘semantickeyphrasetable' TSQL procedure works on a table that consists of Text columns analyses them and returns the key phrases found in the specific text column. This function also returns a score, which specifics the statistical significance of the key phrase within the document.
  • Finding the key phrases that makes two documents similar or related to one another, ‘semanticsimilaritydetailstable' TSQL procedure works on a source text column and a matched text column on a table and returns all the key phrases that are common between the two documents mentioned.

With the above functions we can extract the key phrases within the documents as well as find the matching documents for a document of interest.

Mapping Semantic Search with Sentiment Analytics
The following table explains the typical attributes of sentiment analytics tool and how SQL Server with Semantic Search feature can be fit to that purpose.

Sentiment Analytics Requirement

SQL Server Feature Mapping

Natural Language Processing

The full text indexing options that enables semantic search accepts the language as a parameter. As per the documentation most of the languages like German, English, French, Italian, and Russian  are supported.

Text Analytics

The above mentioned options like the extraction of key phrases together with the Full Text Search features provides a solid building block for Text Analytics.

Storage & Indexing

Both the Full Text and Semantic Search indexes are supported with the storage optimization and flexibilities provided by SQL Server in the form of file group and data file placement.

Transformation

The sentiment analytics is not a one-stop process. With the building blocks provided by Semantic Search, multiple iterations of the key phrases and their relevance ranking needs to be transformed to get the results. SQL server options like SSIS and / or TSQL transformation procedures can play a vital role here.

Analytical Models

The process of classifying sentiments require complex algorithms like ‘Naïve Bayes' which have been already supported by SSAS (SQL Server Analysis Services). With the tight integration between SSAS and SQL Server utilizing these pre built algorithms in to sentiment analytics is quite possible.

Performance / Scalability / Big Data

While Sentiment Analytics can be done without Big Data, generally social media, click stream feeds and other unstructured data been the source for sentiment analytics. Technologies like Polybase which provides seamless integration between Hadoop HDFS stored data with SQL server tables will be an easier choice for sentiment analytics.

Also time consuming process like text extraction, indexing have been optimized using scale up capabilities of SQL Server core engine.

Summary
Utilizing sentiment analytics to promote their marketing campaigns and to improve the product offering will be a key project in most of organizations. With the growth of unconventional data sources the integration of text analytics is also an important aspect. While there are other choices, SQL Server with the integration of other supporting products mentioned above can be an effective choice for performing sentiment analytics.

More Stories By Srinivasan Sundara Rajan

Highly passionate about utilizing Digital Technologies to enable next generation enterprise. Believes in enterprise transformation through the Natives (Cloud Native & Mobile Native).