Description: Explores the open vs. closed systems debate with regards to actuarial modeling and exposes the over-simplification. Valuable for anyone who uses models.

Description: Python has become the fastest-growing major programming language in the world. It has surged in popularity over the last decade, following the development of its industry-leading capabilities in data science, machine learning and artificial intelligence, and it is now one of the most globally popular and widely used languages. In this article, I discuss some of the reasons for its popularity and explain why many actuaries and insurers have started using Python.

Description: CMS provides ACOs with monthly claims data for members, including their medical visits and ACO program eligibility. This data can be divided into Membership and Claims tables through the ETL process. However, when calculating PMPY at specific levels of detail using these tables, issues arise due to unrelated dimensions, leading to inaccurate and understated results. This article presents a solution to this problem by utilizing Tableau's data blending capabilities.

Description: Read the April 2017 issue of CompAct published by the Technology Section. Newsletters provide practical information for the working actuary, articles on trends that impact sections and news about section activities.

Description: The article discusses the use of R in prototyping automation procedures. It offers general pros and cons for R and walks through a specific use case, which involves scraping and summarizing data stored in Microsoft Excel.

Description: Letter fron the Chairman.

Description: The barrier between humans and machines continues to blur, but curiously, this may have been prediction thousands of years ago in ancient myths. This article looks to how some of the early legends are becoming manifest in modern technology, particularly in the area of elder care. From automation, to humanoid assistants, to mind clones will the age of technology bring a new utopia?

Description: This article discussed the NEAT family of neural network algorithms. NEAT automatically determines the number of hidden neurons, as well as their structure. Additionally HyperNEAT can scale to very large problem domains.