Email
Password
Remember meForgot password?
    Log in with Twitter

How machine learning is shaking up investing: Interview Special

By Tim Sandle     Nov 15, 2017 in Business
A new machine learning platform from Harvard graduates is shaking up investing. This is presented as the first completely transparent and accountable prediction tool for thousands of stocks.
The tool is called Market Sensei and it is run by the company Expat Inc. The tool uses patented machine learning algorithms to give powerful prediction abilities to both experienced traders and novice investors. The algorithms create a predicted low, high, opening, and closing price for thousands of stocks.
In addition, every stock’s prediction data also includes a graphical representation of Market Sensei’s prediction accuracy for that stock over the near term. Users of the service are presented with a dynamic accuracy score so they can see prediction performance over time.
To understand more about the platform, Digital Journal spoke with Patrick Kwete, the company’s founder.
Digital Journal: Thanks for the interview. How can machine learning help with investing?
Patrick Kwete: With machine learning and artificial intelligence one can analyze way more financial information than previously possible and better assess risk and rewards of investing decisions. Machine learning also opens the door into more insights about investors on habits which can help them be more successful.
DJ: How did you go about developing your algorithms?
Kwete: My algorithms are all novel and in house, not dependent on any framework from other companies. This makes them very unique and more likely to provide a competitive edge.
The first A.I. algorithm I made went on to get a granted patent which recently blocked an IBM patent application for predictive analytics. So, we know that our innovations have merit and are well ahead of the curve.
A stock investor reacts in front of a screen showing stock market movements in a brokerage house in ...
A stock investor reacts in front of a screen showing stock market movements in a brokerage house in Shanghai
Johannes Eisele, AFP/File
DJ: What types of testing did you undertake?
Kwete: We compute accuracy rates for each of our predicted metrics, so we are able to watch out for differences in our performance. Because our analysis is systematic and we keep ongoing accuracy rates we can continuously refine our algorithms to ensure our optimal performance. That is something no human analyst or advisor could do because there are so many unknown variables.
DJ: How accurate are the predictions? Can you quote any examples?
Kwete: We provide transparent accuracy rates for each of the thousands of stocks we cover. And, our accuracy rates include rates for each of the intraday metrics we currently analyze as well as the overall rate. So there is some variability but in many cases our overall rate for previous 9 days is between 60 percent to 80 percent.
60 percent is our key point because then someone following the analysis all week should make money. The nice part, and differentiator, about our platform is that you do not have to take my word for it. And that is by design.
Any subscriber can see our accuracy rate for any stock of interest so they can adapt their strategy based on the accuracy rates they see. This is a sharp contrast, many people follow other stock analysis or news on a daily basis without being able to know the daily accuracy rate, that breeds an environment that has inherent lack of accountability.
Our platform is not trying to convince people to follow our analysis blindly, in fact we are trying to give investors more transparency than ever before to make better trading decisions.
DJ: How far into the future have the predictions extended?
Kwete: Our mobile app shows daily analysis for the next seven days. However, our web platform shows daily analysis of the next thirty days. So depending on people’s trading strategy, they can pick their ideal time frame. Personally, I love to use the daily seven day predictions to know when to enter a position in a stock I’ve been eyeing.
Phillippine stock market board.
Phillippine stock market board.
Katrina.Tuliao
DJ: Is information displayed and captured in real-time?
Kwete: For our stock analysis, we capture available stock information on a daily basis, and update the analysis accordingly each day as well. On the app and website home screen, and the prediction for the current trading day, or the next trading day, depending on the time.
DJ: How is the information generated – is this in digital form?
Kwete: Everything is digital, all analysis is A.I. powered and there is no human interference. This is done to remove any element of bias from analysis on our platform.
DJ: Is the technology aimed at experienced traders?
Kwete: The technology is aimed for both experienced and new traders. We’ve found that many experienced traders that have used other platforms in the past love ours even more because they have the added transparency of being able to see daily accuracy rates for the first time. We also provide a fun intuition trading game and insights so that even the most experienced traders can leverage our platform to be even more successful. And for novice investors, our interface, the information presented is way less intimidating.
DJ: Can those new to investing take advantage of the technology?
Kwete: Yes, bringing more people into investing and increasing financial inclusion is very important to us. Our Tinder-like intuition training games lets you train your stock trading intuition through quizzing yourself on historical charts. It’s probably the most fun way to learn how to trade. And, next week we are rolling out personalized insights based on patented machine learning on your intuition training.
A photographer takes a picture of the graph of the German stock market index DAX at the stock exchan...
A photographer takes a picture of the graph of the German stock market index DAX at the stock exchange in Frankfurt am Main on June 5, 2014
Boris Roessler, DPA/AFP
DJ: What has the take-up been like?
Kwete: We had paying monthly subscribers from the first month. The platform is almost at 6,000 registered investors and it keeps growing faster organically. That implies its analysis may already be playing a role in the investment decisions of over half a billion dollars in investable funds. Earlier this year, we were invited to showcase at Techcrunch disrupt NY’s Fintech startup alley.
DJ: Are there any other projects that you’re working on?
Kwete: Our goal is to make the benefits of A.I. to all in a manner that positively impacts their daily lives by augmenting people (not replacing them). Our granted and pending A.I patents covers personalized investing, personalized medicine, augmentation of weapons with A.I and personalized education. Each of these have the potential to positively impact many. At the moment, because of the important role the financial sector plays in all our lives, our priority and focus is our financial platform.
That’s why we’re leveraging all our innovations into having a reliable and trustworthy financial analysis platform. We know our continued success in finance will drive our success in these other areas in the future.
More about machine learning, Investigation, Stocks & trading, smart investments, Trading
More news from