3 ms·
Holy crap this is powerful. Is the company simlarity done using a clustering measurement on the corpus of language in the patents? I can't think of a product ma
by motohagiography 5y ago
Holy crap this is powerful. Is the company simlarity done using a clustering measurement on the corpus of language in the patents? I can't think of a product manager whose job does will not depend on this.
- l7l 5y agoHi motohagiography, thanks a lot for the kind feedback, much appreciated. I built profiles for each company based on their filing strategy and technology and then in a second step computed their similarity using Pearson. If you are interested in the details, I put together some of the details here https://medium.com/goodip/ding-the-pearson-correlation-matrix-on-huge-datasets-in-python-bb9680245259 https://medium.com/goodip/ding-the-pearson-correlation-matri... Cheers Linus