DOCOsoft’s Bernard Cosgrave helps develop ground-breaking data analytics algorithm
A learned paper co-authored by DOCOsoft data scientist Dr Bernard Cosgrave, Professor Ali Hadi of Cornell University and Dr Ayman Taha and Dr Susan McKeever, both of Technical University Dublin, has been accepted for publication by top-tier academic journal Expert Systems with Applications.
The paper, entitled A Multiple-Association-Based Unsupervised Feature Selection Algorithm for Mixed Data Sets, is the product of extensive research carried out in partnership with CeADAR, Ireland’s Centre for Applied AI, Technical University Dublin and Cornell’s Data Science faculty.
This has resulted in the development of an algorithm more capable than any previously created of separating out interesting features from within categorical and mixed data sets – thus making them susceptible to advanced data analytics and machine learning – without the unfeasibly long processing times that come with the inclusion of large amounts of irrelevant or redundant information.
Insurance data sets are a classic example of mixed data – i.e. data that combines numerical and categorical information. Developing an algorithm with state-of-the-art performance for clustering and classification accuracy, and better-than-state-of-the-art performance for processing time, is a significant step forward in the field data analytics and machine learning as they apply to the insurance sector.
Congratulations to Bernard, and to Dr Ayman Taha, Dr Susan McKeever and Professor Hadi. We will be publishing further details once the paper has been published.