The use of unparliamentary language at road shows, public rallies, television appearances or posting derogatory comments against rivals on social media platforms, such as Facebook and Twitter, will now cost politicos dear.
In a bid to detect cases of violation of moral code of conduct using technology, the Election Commission, in collaboration with the Centre for Development of Advanced Computing (C-DAC), has developed a unique tool that will screen information and generate reports. This system will scan messages appearing across mediums, such as news channels, newspapers, web portals and social media platforms, for suspected violations.
The system has already been implemented and is adept at recognising Indian languages. Being a pilot project for general elections, the system has been currently implemented in Madhya Pradesh.
"There are several incidents of violation of model code of conduct, which go unnoticed during the elections. So, there needs to be a system, which will keep track of the identified textual data from newspaper websites, politicians' and parties' websites and social media for violations by members or the candidates," said Prof Rajat Moona, director general, C-DAC.
He said the information retrieval system for the Election Commission will try to make relevant data available from newspaper websites, social media and news channels in identifying suspected violations.
The system provides an integrated environment for data mining, text analytics, optimisation and more. It incorporates a variety of Natural Language Processing techniques and processes for the collection, classification, analysis and interpretation of mass media and social media data to retrieve information, leading, ultimately, to new insights.
Dr Hemant Darbari, executive director, C-DAC, said once the suspected data is retrieved from various mediums, the tool will generate a report in English and forward it to the observers of the election commission. The observers can then assess its content to check if there was an actual violation of the code of conduct.
He said that in order to make it intelligent, keywords such as liquor, money exchange and booth capturing among others are fed into the system to trace content.
"It acts like a search engine, where keywords are fed into the system to view the information. It is then filtered further to ascertain, whether the information analysed pertains to poll code violation for the parliamentary elections of 2014 only and not of previous period," Darbari said.
Some unique features
Monitoring and analysis of data available in a digitised form
Date sourced from e-newspapers, websites and blogs of candidates
Monitoring and analysis of identified domain-related data from Twitter, Facebook
Facility for free text search through keyboard input in English and Hindi
Results obtained provided in clustered format for ease in navigation
Bookmark facility on searched video result
Report generation based on results