Most recent
-
Reproducible Research
This article summarizes use of programming collaboration tools and source and version control (SVC) to make research products reproducible. -
Deep Learning in Segregated Fund Valuation: Part 2
This article is the second part of an article that appeared in April 2022 on the Emerging Topics Community webpage. It will discuss the data preparation, hyperparameter tuning and selection, and the training and testing process of the deep learning models. To reach the final conclusions, the article will continue to compare the projected cash flow results from LSTM and LSTM-Attn with those from the traditional method, and evaluate the time series generations of interest rates and equity returns by WGAN and TCN-GAN -
The Probability Principle of Group Testing: The Full-Scale Nucleic Acid Testing in Tianjin
On January 9 2022, a full-scale nucleic acid testing in Tianjin was launched. Over 10 millions of people were tested with the results announced within 2 days. The speedy efficiency was partly due to group testing with 10 persons per group. With this background, the aim of this article is to explain the probabilistic principle underlying group testing. To make the expository vivid, some numerical results and figures were provided using R language, a popular software in actuarial science and statistics. -
Anders vs. Shea, Part 4: A Champion is Crowned
Shea Parkes, FSA, MAAA, and Anders Larson, FSA, MAAA, reveal the results of the competition and share some final thoughts on the 2021 Milliman Health Practice Hackathon. -
Section Elections: How to Become an SOA Section Council Member
Now is the time to become a candidate for an SOA Section Council. Current Section Council members Joe Alaimo and Kevin Durand walk you through why and how to volunteer for this important leadership position. -
Anders vs. Shea, Part 2: Anders’ Story
Shea Parkes, FSA, MAAA, and Anders Larson, FSA, MAAA, are joined by Nick Vander Heyden to discuss the approach used by Anders’ team in the 2021 Milliman Health Practice Hackathon. -
Deep Learning in Segregated Fund Valuation: Part I
Segregated Fund is a special investment fund to provide capital appreciation with embedded insurance features. The traditional methodology to estimate the capital reserve and pricing of contracts goes to Monte-Carlo based stochastic models due to its complexity. Recent research has introduced a deep learning model, Long Short-Term Memory (LSTM), to help cash flow projection for a Segregated Fund in its whole lifetime horizon. In this paper three new deep learning models are presented: Long Short-Term Memory with attention (LSTM-ATTN) to estimate the liability reserve and pricing Segregated Fund contracts, Wasserstein Generative Adversarial Network (WGAN) for stock return forecasting and Temporal Convolutional Network on GAN (TCN-GAN) for interest rate time series generation. As an example, the cash flow projection and Economic Capital for Segregated Fund portfolio are used to compare deep learning models against traditional models in terms of accuracy and computation efficiency. -
Emerging Topics Community: Anders vs. Shea, Part 1: Setting the Stage
Shea Parkes, FSA, MAAA, and Anders Larson, FSA, MAAA, are are joined by the organizers of the 2021 Milliman Health Practice Hackathon: Riley Heckel, FSA, MAAA, Austin Barrington, FSA, MAAA, and Phil Ellenberg. -
Transforming Group Underwriting Using Artificial Intelligence and Machine Learning
A review of opportunities for using machine learning and artificial intelligence in group health underwriting including cost predictors and predicting behavior and decisions. -
Reflecting on the 2021 “Introduction to Modeling Bootcamp” and Changes for 2022
In July 2021, the SOA Modeling Section ran an “Introduction to Modeling Bootcamp” aimed at beginner modelers and students. This article discusses last year’s bootcamp, anticipated changes for this year’s bootcamp, and how you can get involved!
Welcome to our medium.
Dive into insightful content. Gain practical knowledge. Explore the latest research and key information for future use. Discover the Emerging Topics Community, an online forum that focuses on three main topic areas–Modeling, Predictive Analytics and Futurism, and Technology.