DataHour: An Introduction to Machine Learning with Sequence Data

Sequence Data is one that contains ordered elements. For example - the monthly sales over the last year, codes emitted by IoT devices, words in a sentence, etc. Elements forming a sequence can be numbers (eg. sales) or categories (eg. words). The former is generally framed as a time series forecasting problem and most developments in the latter were due to advances in NLP. When the elements are categorical but not words in natural language (eg. codes emitted by IoT devices), it becomes difficult to make sense of sequences. So how do we build systems to learn from categorical sequences?

 In this Datahour, Akshit will cover the following topics in details:

• Introduction to sequence data
• Machine Learning with sequences
• Learning from categorical sequences
• Python code demo

Prerequisites: Interest in Data Science.

Akshit Bhalla is a seasoned Data Scientist, currently working with the CSSD R&A - Innovation team at HP Inc in Bengaluru. With over 3 years of experience in the field, he is well-versed in working with large tabular datasets and sequence data in the form of signals and text. Akshit has demonstrated expertise in domains such as Supply Chain Management and Customer Support, leveraging his proficiency in Big Data and AWS / Google Cloud based MLOps platforms to drive innovative solutions. He possesses a strong skill set in building Deep Learning models and has been instrumental in driving successful business outcomes at HP.