Sometimes, two things just go together, such as peanut butter and jelly or, in the case of Boston-based startup Predilytics, machine learning and health care. The company announced on Tuesday afternoon it has closed a $6 million Series A round with investment from Flybridge Capital Partners, Highland Capital Partners and Google Ventures. It’s not the first application of big data to health care, and it certainly won’t be the last, but it’s application of machine learning to health care providers’ administrative data might be unique.
As we’ve reported before, health care is a major focus for big data companies and data scientists because there’s so much data involved and the problems are so heavy. The right analytic tools could end up saving lives or saving billions of dollars in an industry where just about everyone agrees that costs are out of control. Details are sparse on what, exactly, Predilytics is up to, although it appears to be focusing on the latter type of saving by making health care stakeholders from hospitals to insurance providers run more efficiently and generate more revenue.
Technology-wise, it claims the following:
“Predilytics offers a new approach for generating healthcare insights – applying big data, machine learning technology to create transparent, unbiased, business driven results. This approach provides exceptional predictive models that are 2x to 4x more insightful and actionable than conventional statistical/regression modeling and rules-based methods.
“Our solutions use a full range of available data sources; including administrative data (e.g. claims, Rx, lab, eligibility), service operations (e.g. call center), care management operations (e.g. HRA) and electronic medical records. We can utilize both structured and unstructured data using our natural language data analysis capabilities.”
In English, that means customer data will be analyzed against a system that constantly detects new, hidden patterns and, therefore, new insights without constant human intervention to develop new algorithms or tweak existing models. Because Predilytics leverages natural-language processing technology, users don’t have to mold their data into a standard format or lexicon before the system can make sense of it.
According to a press release announcing the funding, the company “will use the funding to further expand its product offerings and grow operations with a focus on analytics, information technology, application development and account management.”
Predilytics looks to be among a group of early companies building products atop once-inaccessible machine learning libraries. Where Predilytics and other companies such as BloomReach, DataPop and even Bookt apply machine learning to specific use cases, others such as 0xdata, Skytree and WibiData are trying to make machine learning accessible for a general audience.
Feature image courtesy of Shutterstock user sergojpg.