If Dinosaurs Had Big Data, Would They Be Extinct?
A treasure trove of valuable customer data flows through any number of social networks every minute of every day. Making sense of that data, however, is a whole other matter.
To address that issue, DataSift is applying machine learning software to social media data to help take those data streams and create actual actionable intelligence.
Tim Barker, chief product officer for DataSift, says that rather than hiring data scientists to manually classify social media data, VEDO Intent uses a set of Active Learning algorithms to observe how posts are manually classified into categories such as rant, rave, purchase intent or churn. Barker says VEDO Intent then dynamically builds a model to first suggest, and then fully automate, the real-time classification of millions of posts. Access to those posts is provided via VEDO, a programmable data processing engine that via a single application programming interface (API) enables organizations to harness hundreds of social media data feeds.
To make analytics more accessible, DataSift has also launched its FORUM initiative through which its customers and partners can develop and productize advanced algorithms and applications. To facilitate the process, DataSift is promising to publish and maintain foundational statistical models for FORUM members, including the statistical Keyword Relationship Model, available now, which helps marketers identify the keywords and hashtags they should be using in their campaigns.
Machine learning makes the most sense when the sheer volume of data that needs to be analyzed simply overwhelms the capacity of the human mind to process it. In this case, machine learning isn’t necessarily replacing a task that would be assigned to a human being, but rather making it possible to perform an analytics task that simply would not have been practical to achieve any other way.
