A British company called Sharpe Capital is taking behavioral investing to a whole new level. Sharpe plans to use artificial intelligence (AI) to try and learn what investors are thinking.
The goal at Sharpe is to fix one of the biggest shortcomings in AI-based investing. AIs are very good at technical or fundamentals analysis but are unable to judge market sentiment which is based on human emotions and beliefs.
That means an AI can make predictions based on factors such as price, market cap, market volume, and earnings per share (EPS). Yet, AIs and other algorithms often fail to predict market movements, because they are often triggered by emotional factors; such as investors’ or traders’ fear or euphoria.
Spotting the Patterns of Market Sentiment
Sharpe Capital hopes to correct that by teaching AI how to spot patterns of behavior and sentiment in the market, a whitepaper indicates.
That might be easier to do than you think because artificial intelligence is very good at spotting and analyzing patterns. The challenge is figuring out how to locate and discern patterns of market sentiment.
The Sharpe team is trying to discern patterns by setting up an Ethereum-driven platform that harvests market sentiment. Starting on December 11, 2017, the Sharpe Platform is supposed to reward holders of its initial cryptocurrency offering (ICO) SHP that provide opinions on stocks and blockchain assets, with small amounts of Ether.
The amount of data is supposed to be increased by launching a mobile app designed to make it even easier to provide investment opinions. The mobile app is supposed to be launched in 3rd Quarter 2018 after the Alpha Testing of the Sharpe Platform is complete. Beta Testing of the platform is scheduled to begin on 11 December 2017.
Harvesting and Analyzing Market Sentiment for Fun and Profit
The whole idea is to generate vast amounts of “crowd-sourced asset sentiment” for the AIs to analyze.
The market sentiment will be stored in the Ethereum blockchain. Sharpe hopes to locate reliable sources of market sentiment by giving each provider a reputation score. Those with the best reputation scores will receive the highest rewards, to encourage them to offer opinions more often.
Sharpe’s AI will analyze that data and combine it with fundamental analysis is an attempt to predict future stock and cryptocurrency prices. That data will be used to make investments, which are supposed to form the basis of an equity fund. Long-term plans call for the company to start issuing Sharpe Crypto-Derivatives (SCDs) which will pay dividends based on the equity fund’s performance.
That means SCD owners will literally be investing in market sentiment if Sharpe’s plans come to fruition. To reach that goal the company will have to clear a lot of hurdles including regulators, technological challenges, and investor acceptance.
An AI that can Read Market Sentiment
Sharpe has already racked up one impressive accomplishment; building an AI that is supposed to read market sentiment.
This was done by using artificial neural networks (ANN), algorithms that are supposed to mimic human thought processes, and machine learning. Sharpe has used a Machine Learning modeling technique called Natural Language Processing (NLP).
An AI using NLP analyzes patterns of writing and identifies the differences between human-written and computer-generated content. Such an algorithm should also be able to use NLP to spot individual the work of individual writers or posters to social media.
The hope at Sharpe is to be able to figure out what traders, investors, fund managers, and other market players are saying about the market or individual investments. If large numbers of people start complaining about or bad-mouthing a specific cryptocurrency or a stock that might indicate a selloff is imminent enabling Sharpe to short that stock or altcoin. Conversely, large amounts of positive chatter about a small-cap stock might indicate a price increase is imminent.
To work, Sharpe’s AI will have to be able to screen out PR campaigns, false information disseminated by the pump and dump scammers, meaningless opinions, fake news, propaganda, and all the other garbage that’s out there on the cloud. The long-term goal would be an AI that scans social media, news outlets, the blogosphere, professional analysis and many other sources of information in its search for patterns of market sentiment.
Crowdsourced Sentiment Modeling in Use Elsewhere
Such crowdsourced sentiment modeling has already been used for some other applications such as gauging voters’ opinions about candidates or issues. There have been claims that the Trump campaign used such modeling in last year’s U.S. presidential election.
The story is that Donald J. Trump’s son-in-law Jared Kushner hired a British AI company called Cambridge Analyctia to gather data that was used to evaluate social media trends. Cambridge is associated with the historic British university of the same name.
A potentially lucrative use of crowdsourced market sentiment modeling might include attempts to gauge the effectiveness of advertising or marketing campaigns. Another would be to discern consumers’ attitudes about products or brands.
Spotting changes to popular opinions about products or advertising campaigns might save brands vast amounts of money. This might help organizations identify new sales opportunities or unfilled customer needs and demands.
Even intelligence agencies can use sentiment modeling to spot changes in behavior or attitudes among terrorists or enemy soldiers. An AI might be able to monitor chatter among enemy troops; and discern if they are becoming disillusioned, and getting ready to surrender or quit the fight for instance. It might almost be able to predict a revolt or rioting, by modeling the level of rage average people are expressing on social media.
Sharpe’s Technology is a Potential Gamechanger
Sharpe’s technology is a potential game-changer; because its engineers may have figured out how to build a cheaper AI using low-cost cloud architecture, provided by Amazon Web Services (AWS). That means Sharpe might be able to build lots of AIs fast at a low cost using the auto-scaling groups and server-less architecture available Amazon’s (NASDAQ: AMZN) AWS Lambda.
If such claims are true, it might be possible to build powerful next-generation AIs capable of sentiment market modeling almost anywhere a good internet connection and decent computers are available. All that will be needed are capable engineers or AI-building AIs, which Alphabet (NASDAQ: GOOG) the company formerly known as Google is trying to create.
This might be the beginning or symptom of a paradigm shift in AI everybody should pay attention to. What’s really exciting is that almost anybody might be able to invest in it and participate in the Sharpe Platform.
Sharpe Capital ICO Begins November 13
Sharpe Capital is holding a presale that is scheduled to end on 13 November 2017. The initial cryptocurrency offering (ICO) for the SHP Tokens is scheduled for 13 November 2017, (Monday) at 2 P.M. Greenwich Mean Time (GMT or London Time).
For more information about the ICO please go to https://sharpe.capital/tokens. You will have to buy Ethereum and use the MetaMask or MyEtherWallet applications to participate in Sharpe Capital’s platform.
Those interested in a highly speculative and extremely risky investment with a potentially big payoff should checkout Sharpe Capital. Even if you do not invest you should research Sharpe because its plans are showing us how the market will probably operate in a few short years. Therefore every investor needs to be aware of market-sentiment modeling and understand how it is supposed to work.
Blogger’s note: some of the text in this article has been changed from the situation to more accurately reflect the situation. Please note that reflects a change of perspective and not a correction.