Frequent Pattern Mining: Real Life Examples

Frequent Pattern Mining: Real Life Examples
What is frequent pattern mining? 

The Benefits
Workers Compensation is an area of that is open to the benefits of frequent pattern mining, particularly when providing insurance to those businesses with repetitive manual tasks, according to a white paper titled Mine Your Business—A Novel Application of Association Rules for Insurance Claims Analytics. The paper reveals that eye injuries are highly associated with day shifts, and the hazard of being hit by a foreign body. Of all the claims associated with day shift and injury by a foreign body, 81% result in eye injury. This suggests that the company will benefit from implementing and improving the eye safety program for its employees.

Costly Claims
Fracture claims are another area where patterns using a rule can help to mitigate potential harms. This rule tells us that 80% of the lower leg injuries for day-shift drivers result in fractures. The analysis of claim data shows that fracture claims are some of the costliest on a per- claim basis. This suggests that a company would benefit from implementing safety programs to prevent lower-leg injuries.
 
Claims Don’t Like Mondays?
The final rule tells us that 40% night-shift injuries occur on Mondays. On the face of it, this rule does not provide a clear explanation of what is going on here and does not suggest a course of action. Knowledge of the business may reveal the reason behind this pattern. Perhaps a large number of night-shifts occur on Mondays, after the weekend parties at both businesses and homes. So perhaps this rule is just reflecting this pattern of how the night-shifts are distributed during the week.

Helpful to Insurers
As the paper mentioned above explains, 60% of Workers Compensation claim dollars arise from 5% of the claims for some self-insurers. Even a relatively small number of such high-loss claims can result in a disproportionately high cost to the insurer. Clearly, any methodology that provides insights into the characteristics of such injuries and resulting claims will be immensely helpful.

Rich Data
DOCOsoft is investing a lot of time and resources in advanced machine learning.For us it’s a scientific endeavour but it’s also a business process transformation as well, because we can see that the Lloyd’s building is closed, that people are working from home so change is happening. We can cope with that but we can also see in future the importance of machine learning so we’re also helping our clients to prepare for that.

Our research will enable us to use rich data to be able to predict, say, reserve value to find interesting frequent patterns within the data. So, if you’d like to cooperate with us on our research on machine learning supported by our CMS, we’d be happy to talk to you.