Time Series Analysis and Forecasting in Java

6:30 - Food and drinks
7:00 - Time Series Analysis and Forecasting in Java: Motivations, Solutions, and Current Challenges
8:00 - End
More food and drinks and chat!

Abstract:

In this presentation, Jacob will walk through his unexpected journey of going from a recent college graduate with no Java or time series analysis experience to writing a pure Java implementation that included, among other things, seasonal ARIMA modeling, exponential smoothing state space models, and seasonal trend decomposition using Loess. He will also talk his about his motivations for writing an open source time series library where previously none existed, the challenges presented there, and future directions of that project.

About the Speaker:

Jacob Rachiele is a statistical software developer with expertise in combining Java software with classical time series analysis and forecasting.
He was the first developer to write an open-source Java implementation of Seasonal ARIMA modeling and forecasting. At Baxter Planning, where he was worked the last four years,
he also wrote a pure Java solution to time series forecasting that included an implementation of Rob Hyndman and Anne Koehler's exponential smoothing with state space models, along with Croston's intermittent demand method, season trend decomposition using Loess (the STL function in R) and automatic model selection based on goodness of fit criteria. He is currently interested in finding the best ways to incorporate time series analysis and forecasting with large scale streaming data: think R's forecasting package combined with reactive streams.

Bring as many guests as you want.

To view our Code of Conduct and more information about our meetup, visit our site (https://www.influxdata.com/timeseriesmeetup/code-of-conduct/)