Lifestyle Article - How to read 18,701 documents in less than 2 hours


We read 18,701 job adverts to answer the question: what does an actuary do? Reading this many adverts should have taken weeks, so how were we able to do it in 2 hours? Machine Learning! We read the data into R and let a machine learning algorithm called “Topic Modelling” do the hard work. What we found was very surprising….

 The algorithm identified 40 different “topics”. We can explore the meaning of each topic using an interactive visualisation – to start interacting click here.
 

 Clicking on bubble 32 allows us to reveal the most relevant words for that topic which are displayed below. This topic is best described as “Investment & ALM”.

 

 Have a play with the visualisation and see how many topics you can identify.

 What does an actuary do?
 The algorithm automatically identified the major types of Actuarial work including Speciality Pricing (5) Capital Modelling (6/29) Pensions (10) Annuity Pricing (11), Reserving (13), Reporting (15) Personal Lines Pricing (16) Risk Management (18) Life (19/24) Prophet Modelling (21) Solvency ii (25) Catastrophe Modelling (27) Investment and ALM (32) and Marine/Aviation (35). The algorithm also detected softer skills such as Project Management (1) Leadership (2) as well as experience level such as Junior(36), Making Progress through exams (3), Newly/Nearly Qualified(20) and Seniority (7).

 Expect the Unexpected
 Having delivered several unstructured data projects in different industries one learning point stands out – expect the unexpected! Unexpected results include
 • Online Casino Chat Logs – “We didn’t realise customers were having difficulty depositing with their credit card!”
 • Insurance Company Loss Descriptions – “We shouldn’t be paying claims for that event– it’s not covered by the policy!”

 Still we were surprised when we clicked on topic 26 to see a topic dedicated to French actuarial roles! Furthermore when we investigated the topics in the top right of the bubble chart we were fascinated to see they related to the various actuarial recruitment companies! The algorithm had managed to automatically differentiate between the role requirements and the recruiters “sales pitch”.
 
 What did we learn?
 Since the role of an actuary is well understood, we have essentially validated our machine learning approach as a method of discovering new and reliable patterns from new, unexplored data items. The next time we want to quickly understand a large collection of text documents we should consider using this approach. To learn more about this fascinating subject click here

 Machine Learning Solutions provide training to actuaries in Machine Learning techniques as well as delivery of technical projects in predictive analytics across various industries. For more information visit here 
  

Back to Index


Similar News to this Story

Paul Macro joins The Aon MasterTrust board
Aon has announced that Paul Macro is joining the board of The Aon MasterTrust. He replaces Jonathan Reynolds who is stepping down from the board after
Nikhil Rathi reappointed as FCA Chief Executive
HM Treasury has today announced the reappointment of Nikhil Rathi as Chief Executive of the Financial Conduct Authority (FCA) for a second term. 
Just Group appoints new Managing Director of retail
Just Group plc (“Just”), the FTSE 250 retirement specialist, has appointed Tom Evans as the Managing Director of its Retail business, effective from 1

Site Search

Exact   Any  

Latest Actuarial Jobs

Actuarial Login

Email
Password
 Jobseeker    Client
Reminder Logon

APA Sponsors

Actuarial Jobs & News Feeds

Jobs RSS News RSS

WikiActuary

Be the first to contribute to our definitive actuarial reference forum. Built by actuaries for actuaries.