The algorithm is private, but these sites function in a straight forward manner. Think of it as psychology coding 101.
The easiest way to emulate is this process is through a numerical scale. Take into account social indicators such as personality type (A,B), age, affluence, beliefs, values, perspectives, and most importantly what you are looking for in the other person. These scores are then "coded" by attributing a number to each answer which is then added or subtracted depending upon response. These values ultimately create a number and that number is "matched" with similar numbers (matrix matching) and you can layer this process to be more precise.
In terms of your project: trying to pair up entrepreneurial minded people. I would start with a specific geographic location, Say... the Bay area. Right there the parameters are defined to specific coordinates (numeric). I would rate everyone on a proximity rating with each other first (which i am assuming that you are trying to do with the 1 city limitation) people that are with in 10 mile of each other receive a rating of (10), 15 miles (9)... 50+ miles (1) ect. The logic that drives this the proximity principle: objects that are close together will be seen as belonging together, so people that live in close proximity to each other will perceive the other to belong.
Next I would create an in-depth survey to try to generate as much relevant information such as skill sets, interests, beliefs, values, age, what they are looking for in a "match" and also they type of company that they are trying to build. If it is extremely technical then there should be 2 very technical co-founders, if less technical then I would say 1 technical & 1 semi-technical co-founders. Depending upon what needs the project has. I would then match people who have "complimentary" scores.
I would then generate scores using my coded survey and match the people together.
In terms of your project, I think it would be a good idea to go through different online survey tests (match, zoosk, eharmony) to see how the questions are set up and then coded.
If you have not done so already I would look at the algorithms that E-harmony, Zoosk and Match.com use to link people up. Because most co-founders will tell you that when they entered into a business partnership it was like entering into a relationship as well. Additionally if you are looking to match start-up co-founders together I think it would be better to focus on close proximity geographic locations, such as Bay Area, LA basin, San Diego.
From one college student to another, if you can find any internships for start ups in your area do it. If you can't then get involve into any business our entrepreneurial clubs. The people in those organizations are usually are looking to engage themselves in business oriented projects. Be willing to learn as much as possible.
Go to your student entrepreneur clubs, business clubs, computer clubs, or other business oriented organization at your school. You may be surprised and find that what you are looking for it right in your area.
Dilution is apart of the entrepreneurial game, I am not surprised that Westergren only has 3%, his value proposition for the first 2 years to investors was less than stellar. investors that put money into this type of startup require large amounts of equity because they need assurance that they will receive a return on their money that they pay forward
For all the applicants on the fence my only response is that yo have nothing to lose. If you don't get picked then so be it but there have be a lot of successful start-ups that are not chosen to participate every year. If you are placing all of your chips on the YC to help you make it then I think it would behoove you to reassess your goals. start ups are about creating something that solves a problem, changes lives and to make okay products better. The initial start up should be about creating a stellar product more than it is about money. To those critics, yes money is a crucial aspect of business but start-up have the unique ability to bootstrap and work under financial constraints that most would never be possible. Creating a start up is as much about the business as it is the experience of starting that business. So go for it you have nothing to loose.
The easiest way to emulate is this process is through a numerical scale. Take into account social indicators such as personality type (A,B), age, affluence, beliefs, values, perspectives, and most importantly what you are looking for in the other person. These scores are then "coded" by attributing a number to each answer which is then added or subtracted depending upon response. These values ultimately create a number and that number is "matched" with similar numbers (matrix matching) and you can layer this process to be more precise.
In terms of your project: trying to pair up entrepreneurial minded people. I would start with a specific geographic location, Say... the Bay area. Right there the parameters are defined to specific coordinates (numeric). I would rate everyone on a proximity rating with each other first (which i am assuming that you are trying to do with the 1 city limitation) people that are with in 10 mile of each other receive a rating of (10), 15 miles (9)... 50+ miles (1) ect. The logic that drives this the proximity principle: objects that are close together will be seen as belonging together, so people that live in close proximity to each other will perceive the other to belong. Next I would create an in-depth survey to try to generate as much relevant information such as skill sets, interests, beliefs, values, age, what they are looking for in a "match" and also they type of company that they are trying to build. If it is extremely technical then there should be 2 very technical co-founders, if less technical then I would say 1 technical & 1 semi-technical co-founders. Depending upon what needs the project has. I would then match people who have "complimentary" scores. I would then generate scores using my coded survey and match the people together.
In terms of your project, I think it would be a good idea to go through different online survey tests (match, zoosk, eharmony) to see how the questions are set up and then coded.
(Here is a link to good coding practices. I think it will be able to help you further understand how you generate a numerical score off a survey) http://books.google.com/books?id=dB0fw5lf0GQC&pg=PA345&#...