Time series forecasting using ARIMA
Details
(Please, ensure that register to this event at our Eventbrite page
https://www.eventbrite.co.uk/e/time-series-forecasting-using-arima-tickets-74863220909, Meetup registration alone won't be valid!).
In this workshop, we will be reviewing the basic statistical theory and a straightforward worked example of Autoregressive Integrated Moving Average (ARIMA) models for time series analysis. As one of several alternative approaches to forecasting future trends using past information, ARIMA is a flexible but also powerful ‘semi-parametric’ method for forecasting and fitting longitudinal (aka panel) data, and there is also a user-friendly R implementation. In contrast to more inferential tools like Cox survival models, ARIMA models can make useful, potentially more accurate forward predictions with relatively sparse datasets, and they can be very useful for surfacing initial evidence of causal trends for further validation.
ARIMA can be used for everything from forecasting highly seasonal sales trends for food products, to the main example we’ll work through: surfacing a potential link between delayed discharges of hospitalised NHS patients and the unusual increase in English mortality rates since 2015.
The workshop does not presuppose any prior experience with time series/panel data analysis, but it will be helpful to have some statistical background as well as basic R skills. Bring your laptops! Data and code will be circulated on Github.
The talk will be given by Jules Shen, who is a statistical consultant with the Health Foundation, working primarily with the Improvement Analytics Unit. She is also a part-time PhD student in public health and policy at the LSHTM, who previously worked in development/health economics and management consultancy. Jules’ wife has worked out that her optimal day is a generalised linear function of pet budgies + ramen noodles.
Agenda:
6pm - snacks and networking for R-Ladies and hosts [lower ground floor downstairs; LSHTM refectory]
6:30ish pm – workshop starts! [upstairs in 1st floor John Snow lecture theatre, please note that no food is allowed in the lecture theatre]
8:00ish pm - end of workshop!