Munich (ots) - - Launch of BMW Group Dialogues 2018 in Los Angeles - Main focus: better mobility for better cities ...
Social Media Monitoring: Automated Relevance Recognition with Artificial Intelligence
Breakthrough in the reliable identification of relevant social media posts
Munich (ots) - Based on its constantly developed technology and specific methodology, German company Consline AG has made a breakthrough in automated recognition of relevant posts. The basis are Deep Learning techniques of Artificial Intelligence (Naive Bayes, SVM, SGD), which, like the human mind, can improve their performance in distinguishing relevant and irrelevant through exemplary learning.
If you google you know the issue: Demanding search queries which go beyond weather or shopping result in countless hits, which still have to be read and filtered by relevance. Likewise, common social media software solutions collect numerous results (so-called mentions), which are irrelevant or have nothing to do with the topic at all. This further distorts sentiment recognition (identification of a positive or negative opinion), which is not very reliable anyway. The keyword-, rule- and/or semantic-based index methodology used by typical social media software cannot reliably distinguish relevant from irrelevant hits.
The AI method developed by Consline goes far beyond keywords, weighting of previous search queries or filters. Instead, all social media posts are processed by the AI software and the complete content, the text structure as well as all meta information (source, author, links, time, etc.) are evaluated. Thus, relevant contributions are identified with >95% reliability. As the remaining irrelevant posts are manually corrected, the system constantly learns and improves the results even further. Thus, Consline provides only relevant results to its customers, who depend on the highest reliability e.g. in the field of product monitoring, even in the detection of rare events. The Artificial Intelligence used by Consline is similar to the systematic pattern recognition in autonomous driving, where sensor images are evaluated in context (for example, light/dark distribution, lines, shapes, colors, etc.) and identified through comparison with learned images.
Consline AG (www.consline.com) is a pioneer in the field of complete and precise monitoring of customer voices and company information online. Since 1999 Consline has been supporting companies in improving products, services and campaigns as well as legally compliant product monitoring. The CIMS (Consline Intelligence Management System) developed by Consline is based on a unique combination of state-of-the-art web technology, qualified employees and international industry knowledge.