Sorry if this is too easy.<p>Image you have all the public LinkedIn profiles :)<p>How would you write an algorithm that:
* receives an input of linkedin profile(s)
* returns an output of different linkedin profiles<p>such that output profiles are 'similar' to input profiles<p>Example:
- Marketing segmentation analysis. You give me your current customers e.g. head of sales at slack etc and I give you other head of sales at tech companies
- Investors: You give me a list of investors you are currently talking to and we give you back a list of potential investors in your space
- Sales
- Hiring
etc<p>Any ideas or links would be greatly appreciated. Are there any general algorithms that can do that?
Look into recommendation systems. A lot of learning materials out there (check out the Google course and this awesome list[0,1]). Start with some simple stuff like collaborative filtering and content based filtering.<p>0:<a href="https://developers.google.com/machine-learning/recommendation" rel="nofollow">https://developers.google.com/machine-learning/recommendatio...</a>
1: <a href="https://github.com/grahamjenson/list_of_recommender_systems" rel="nofollow">https://github.com/grahamjenson/list_of_recommender_systems</a>
If I were in your place, I would start by scraping LinkedIn using Python or Golang and creating a large database, keeping relevant information and keywords for each user. Then I would come up with some sort of clever algorithm to sort the data into some kind of sparse graph with takes into account the distance between the nodes (i.e. how close each of the user is to one another) and then use a search algorithm to search for a specific user and or a number of closely related users.