Description:

sktime is the most widely used scikit-learn compatible framework library for learning with time series. sktime is maintained by a neutral, non-profit under a permissive license, easily extensible by anyone, and interoperable with the python data science stack.

This workshop introduces the usage of sktime for time series-related tasks (forecasting, anomaly/changepoints, classification), and then focuses on a deep dive into Deep Learning algorithms, interoperability with Deep Learning backends, and Foundation models for forecasting.

Deep learning and foundation models in particular are a development focus of multiple ongoing projects, with a focus on providing unified interfaces. We invite the community to contribute and provide feedback.

The workshop will cover the following topics:

• Introduction to sktime toolbox.
• Forecasting, Classification, anomalies/changepoints and time-series analysis with sktime.
• A concise introduction on forecasting pipeline and a tutorial on forecasting with sktime.
• Building different Deep Learning Estimators for time series-related learning tasks.
• Foundation Models, Hugging Face connector, pre-training and fine-tuning

After the workshop, attendees will have acquired familiarity with a broad range of methods for time series in python, including some of the latest advancements in AI for time series.

Prerequisites:

Intermediate Python, Basic understanding of Machine learning and scikit-learn

Content URLs:

Notebooks Introduction Forecasting with sktime Advanced Forecasting Material

Previous Conferences Pydata Global 2021 Pydata Global 2023 Advanced Forecasting : Pydata Berlin

Speaker Info:

Felix Hirwa Nshuti : GSOC contributor at sktime Pranav Prajapati : GSOC contributor at sktime, TFUG Nashik Organizer

Speaker Links:

Felix Hirwa Nshuti * Github * Linkedin

Pranav Prajapati * Github * Linkedin