JCB663 - Tuning Your NLP Model for Market Relevance

Course Description

With your text data effectively cleaned and primed for an algorithm, you're now poised to put it into practical use. While you've created Latent Dirichlet Allocation (LDA) models in prior courses, you've done so using default settings, which may not be ideal for the specific data at hand. To fully ready your models for active portfolio management, you need to train and evaluate them against an industry standard. Only with this assurance can you make associations that are relevant within an investment context, enabling you to construct portfolios of companies that align with a desired industry sector or theme.

In this course, you'll train a variety of LDA topic models in an iterative process to enhance their performance. You'll evaluate their alignment with widely accepted industry classifications to compile lists of comparable companies relevant to a specific investment theme. The process will range from fine-tuning various hyperparameters to optimize the LDA algorithm's learning curve to calculating distance metrics for comparable companies to ascertain their topic similarity with respect to an investment benchmark.

As you progress through the course, you'll conduct an array of comparative analyses to discern the strengths and weaknesses of the LDA approach. Recognizing these aspects is crucial when it comes to the construction and management of investment portfolios. By the end of the course, you'll be adept at training, refining, and applying LDA models, paving the way for smarter, data-driven investment decisions.

The following course is required to be completed before taking this course:

• Preparing Data for Natural Language Processing
• Cleaning Text Data to Optimize Model Performance

Faculty Author

Benefits to the Learner

• Tune hyperparameters to optimize LDA topic model performance
• Compare LDA topic model groupings with an industry standard
• Map LDA topics to companies for portfolio construction

Target Audience

• Financial analysts
• Quant finance investors
• Market analysts and business analysts
• Data scientists
• Software engineers

Applies Towards the Following Certificates

• NLP for Finance : Core Courses