There’s a super cool team that works at the intersection of economics and machine learning. I’m not sure I could get on the team without a PhD in economics or being a ML expert. The downside of the team is you cannot publish your work which is a deal breaker for some PhD holders.
That seems to be exciting work. Are there any companies in the Bay area working on stock trading algorithms, or economy predictions etc?
What is the difference from econometrics? Relabel statistical analysis as machine learning?
It’s next level when you apply machine learning to it. You can input tons of data and machine learning will find the correlations for you. You can analyze far large amounts of data far more efficiently.
I always wonder how to get the ML data. Only large companies with customer bases have it? So, it’s a leg up for them compared to other companies?
I am trying to understand this better. Do you mean unsupervised learning where you don’t specify the features? How frequently does it produce valuable insights?
My understanding is you specify your output, and you can specify a ton of inputs. It’ll determine which inputs are the key driver of the result for you. The simple example used is will you play golf that day. You can give it a ton of inputs, and the machine learning will realize you play golf if you have no afternoon meetings, it’s between 70-80, there’s no rain, and your friend Bob’s schedule is also free. Then it can use that to predict whether or not you’ll play golf in the future.
You can do that for all sorts of consumer behaviors. It also allows you to do small tests ($10 gift card on a specific purchase) and accurately predict which customers will respond to that offer. That way any marketing can be very specifically targeted, so customers see it as relevant and useful vs. annoying.
Do you guys use something like OpenCV or R?
Not good, solution is dependent on your inputs. Can I specify output only?
I have no idea.
They are not specifying what to look for in the inputs. They just provide the whole data. The specifics @marcus335 mentioned (e.g. no rain) are what the algorithm figures out on its own.
I know but it means MLs only look at those inputs which could be totally not influencing the outputs. It also means ML is only as good as the guy who specify what are the inputs to look at. And the quality of the inputs collected.
Our politicians, many of them with liberal arts degrees themselves have long bashed the liberal arts. President Obama once said “I promise you, folks can make a lot more, potentially, with skilled manufacturing or the trades than they might with an art history degree.” Mitt Romney scoffed that “as an English major your options are uh, you better go to graduate school, all right? And find a job from there.”
Their cynicism comes as higher education has turned towards vocational education, i.e. job training. Too many politicians now see college as a training program for under- and unemployed workers on the low end, and for developing STEM (science, technology, engineering and math) workers on the high end. Either way it’s a technical education. And this framing has affected how parents and high school students look at college, as too many see majors as advanced vocational training.
What a bunch of BS. You can teach just about anyone the liberal arts stuff. Entry-level people aren’t asking the questions they mention. Those are questions executives need to ask, and that’s learned through years of experience. Entry-level people need to be able to produce work.
A vast majority of people can’t even get into engineering programs let alone graduate from them. Education should be focused on getting someone a career. If you just want to have knowledge, go read some books. You don’t need to pay $60k/yr at a private university for it.
Liberal arts is bad for the society. Too many liberal arts graduate are the reason for too many lawsuits, and the absurd “femenists don’t eat eggs” animal rights ad
Many engineering graduates are much better in strategic thinking than liberal arts. A few liberal arts are good at office politics and PC and climb the corporate ladder skillfully.