0

I am not sure if this question is on-topic or not: Do you know of any business case where open data machine learning models can be useful? If so, can you point to the open data and the business case it provides?

I am asking this question, because I was thinking about open data and machine learning and how data scientists who create such models can profit from it.
I created a small prototype marketplace for machine learning models to sell and buy pfa, pmml or onnx models: https://machine-learning1.sharetribe.com/infos/about

I thought to pose this question to the data science community so that others can provide some guidance on what can be done with open data and machine learning.

Thanks for you help. Also I am not sure how to tag the question.

|improve this question

migrated from datascience.stackexchange.com Oct 30 '18 at 19:52

This question came from our site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field.

  • how are you maintaining open data licensing/policies with your marketplace? personally, its very off putting to see a question like this here. how will you be giving back to the open data community through your marketplace? – albert♦ Oct 31 '18 at 14:00
  • @albert: since to create the machine learning model, requires some work besides having access to the data, it should be left to the producer and the consumer of the model, what licensing-policies they chose. What do you think about this approach? – orgesleka Oct 31 '18 at 14:03
  • i think you will be violating many licenses and policies by creating a product on top of them. you will need to read them explicitly. i get that there is time put in there. there is also time put into the data you seek access to freely. these communities build atop each other via this symbiosis. there is always room for the marketplace, but nine times out of ten, in my opinion, they are simply leeching off the community. – albert♦ Oct 31 '18 at 14:08
  • @albert: Ok, so what you are saying is that each model producer should use only data which allows this, for example data released under the mit licence? (Thank you for your input) – orgesleka Oct 31 '18 at 14:10
  • 1
    sorry english is not my native language. i did not meant to sound greedy or so. maybe " to benefit" is the better word – orgesleka Oct 31 '18 at 18:10
2

There are various active uses of machine learning in the industry, from computer vision to finance. Let me explain what I do: I currently make a living is that algorithmic trading by ML. My model does not predict the market, it predicts how to play with it. The model can return 10-12% from the capital on FOREX EUR/USD without applying “leverage”.

What it does mean as a business application? Let’s say that you are an investor that has $100,000. A bank will you a give rate, let’s say, 4-5% yearly. My company will offer you 7.5% yearly interest for your money to be in my portfolio. I will trade with my portfolio which I acquired from the investors like you to make a return of %10-12, I will pay you your yearly 7.5% as the company in return where the remaining is our profit.

Some other applications of ML in FinTech:

https://www.kdnuggets.com/2018/05/top-7-data-science-use-cases-finance.html

You may want to check some powerful concepts which are actively has industrial applications via ML; Computer Vision and Natural Language Processing (NLP). Also, for more specific applications, you can refer to:

https://www.kdnuggets.com/2017/01/6-areas-ai-machine-learning.html

|improve this answer
  • thanks for your answer. how do you use open data? – orgesleka Oct 30 '18 at 9:58
  • There are open Python libraries that enables me to pull public FOREX or cryptocurrency data in real-time. For example Poloniex API lets me pull the financial features Close/Open/High/Low/ Volume for cryptocurrencies per 1m/30m/2h/1d, depending on my choice. Then I process that publicly available data for my training purposes. One example would be that. Another example would be that Twitter API allowing us to pull any public tweets with specific keywords or account IDs; lots of companies implement sentiment learning on a specific concept to predict the current hype for a stock, an election etc. – Ugur MULUK Oct 30 '18 at 11:39
  • thanks. very good hint with the sentiment classification – orgesleka Oct 30 '18 at 15:26
0

I found another application:

open data + machine learning = business value

|improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.