Learn how to master the art of feature engineering

Feature engineering is arguably the most important step in the data science workflow. This is why our upcoming webinar 'The Art of Feature Engineering' is fully dedicated to the topic. Over two online sessions, you will learn how to apply common methods and best practices, avoid pitfalls and use feature engineering to successfully fuel machine learning and data mining algorithms.

Feature engineering is the creative process of extracting and selecting relevant, informative and distinguishing characteristics from your data, which can then be used as input for data mining or machine learning algorithms. As the quality of your features largely influences the quality of the results, feature engineering is considered the most important step in the data science workflow.

While it is typically a creative and labour-intensive process, understanding the methodology, tricks of the trade and common pitfalls can help you go a long way. In our next webinar 'The Art of Feature Engineering', part of our Mastercourse ‘Data innovation’ 2021, we will discuss in detail the methodology behind it, including various methods for feature construction, selection, normalization, etc.

Why enroll for our webinar?

Feature engineering is not only one of the most important steps in the data science workflow, it is also the step that will require most of your time. In this trial-and-error process, you will experience that it is possible to extract a wide variety of features from your data. Consequently, you will need to validate which features work and which don't, refine them or define additional ones, validate them again, etc.

Even though feature engineering is an iterative and creative process that comes without an instruction manual, there is a methodology behind it involving some standard approaches, guidelines, recommendations, tricks of the trade and pitfalls, that you should be aware of and that can greatly help you in finding the most relevant features. By registering for our upcoming webinar, you will:

  • Gain deeper insights in the iterative process of feature engineering
  • Learn the tricks of the trade and how to avoid common pitfalls
  • Reach better results when applying a machine learning algorithm

> Register now

Use case: Predicting electricity consumption of a portfolio of households

An electricity provider’s portfolio can consist of many households that consume and produce electricity in different ways. In this research project, the business question was to construct a predictive model that incorporates the potential electricity consumption of individual households, enabling a portfolio manager to have a more accurate idea of the net consumption in his portfolio. The motivation behind this is that it is cheaper to buy electricity on the day-ahead market than on the spot market.

The challenge was to characterize individual households with a set of characteristics or 'features'. This would allow us to cluster them in groups that are similar in terms of energy consumption, enable comparison between different households and train predictive models which can make accurate predictions for the whole group.

After a thorough exploration of the available dataset and based on domain knowledge, it became clear that a diverse set of factors influence energy consumption. This included the day of the week and time of the day with the highest energy consumption, particular weekly patterns such as recurring energy consumption on Wednesday afternoon, etc. These characteristics or ‘features’ could then be used as input for the predictive model.

Learn how to master the art of feature engineering

Would you like to achieve better results when applying intelligent algorithms? Or are you simply interested in learning more about the creative process of characterising data in terms of features? Our two-session webinar 'The art of feature engineering' on 20 and 22 April, is open to any entrepreneur interested in the possibilities data and AI can offer industrial business, including SMEs. You have the option to follow this one session alone or combined with other sessions of the Mastercourse cycle ‘Data Innovation’ 2021.

> Register now

 


#industriepartnerschap #sterkondernemen