PitchBook blog

Your resource for all things PitchBook

PitchBook processes billions of data points to build Suggestions

June 8, 2018

Identifying all the companies in a rapidly changing industry can be a frustrating and time-consuming process. We provide thousands of search criteria to choose from so you can zero in on exactly what you’re looking for—but it's possible to inadvertently filter out the companies you'd like to see or would find interesting. Our newest feature, PitchBook Suggestions, leverages machine learning to help you easily discover relevant companies and fine-tune your search results without having to go back and modify your criteria.

With Suggestions, you can add and remove a suggested company—or populate more similar companies based on a specific suggested company—directly within your search results. This ensures you'll efficiently find all the key players comprising an emerging industry and can also help you find noteworthy new companies.

Where can you find Suggestions?

Under the Companies tab.
Your results will look the same as they always have after running a “Companies & Deals” search in the platform, but now you’ll be able to further customize your view with the “Suggestions” tab. Just look for the yellow plus sign and click to see our recommendations of companies you might want to add to your list.

How do we identify relevant companies?

Three key ways.
Using machine learning technology, your search results, and trending companies, we identify similar companies that may not fit every criterion you set, but may still be worth considering—and adding to your results. The algorithm that powers Suggestions is based on machine learning technology and natural language processing.

All descriptive text within the Platform is encoded into vectors to create PitchBook’s word embedding model. During an advanced search, the Suggestions algorithm encodes the entire search as a vector and compares it against PitchBook's larger word embedding model to locate similar companies. PitchBook’s web mining tools and automated news collection technology has now processed 30 billion words, a 1,000% increase from 2016, strengthening our proprietary word embedding and word vector models to produce high-quality Suggestions.

Want to see Suggestions in action?
 


Not a PitchBook user yet? Request a free trial. 

 

 

Related content