Recently it became known that the Nasdaq is going to introduce its hub a new analytical tool based on machine learning, which will process the data of users of social networks, providing institutional investors a new tool to analyze the market. However, the questions Bitnewstoday.ru attracted by Nasdaq specialists direct answer is not given, only evasively said about the “interesting experience” and then referred to a subscription about nondisclosure. Apparently, it can only mean that developments are tested.
We will remind that earlier in March this year Thomson Reuters launched an improved version of the service MarketPsych Indices, which makes predictions of the market based on more than 2000 news sources and social networking 800. This tool based on AI analysis uses not only quantitative, but also the emotional metrics of traders, i.e. behavioral Economics. Thus, the service promises to its customers to generate more accurate predictions and to choose the best strategy.
Mechanisms and strategies for traders
Usually in its analysis traders use few technical indicators: level of support and resistance, moving average, total trading volume to the relative strength index and the stochastic indicator. Among technical indicators also use trend lines that display information about the characteristic direction of movement of scriptactive to forecast future trends.
However, due to the high volatility of cryptocurrency these trends is very difficult to determine, that is why traders rarely take into account this indicator. But in General, technical analysis of the market today is not sufficient for prediction, since it ignores the political and socio-economic situation in the world.
Mark Lind, Advisor to IBM, which is in the top 15 experts in digital transformation, in an exclusive interview for Bitnewstoday.ru confirmed that from 2017 the use of AI to process the database was for a large company the major trend in the race for profit: “Now the main application of neural networks in Economics is prediction markets, optimization of commodity-money flows, the analysis and generalization of various social surveys, prediction of the dynamics of political ratings, optimizing production process, comprehensive quality diagnostic products and much more. Algorithms based on biology, in particular artificial neural networks and genetic algorithms are the main types used for trade analysis, risk measurement and forecasting of prices.”
All traders use neural networks
To date, the financial and socio-economic system man can not be predicted: too much input. And here comes into the arena AI, able to analyze the different spheres of life and to conduct an independent analysis as well as any financier, while possessing a unique ability – lack of emotion. But then the question arises: why all traders still do not use AI in their forecasts, or whether the neural network is a versatile tool for traders in the near future?
According to Mark, LINDA, “she tells the tech industry, how these technologies can be used to gain a competitive advantage and create new revenue streams where large vendors such as IBM with Watson, use more complex integrated tools to help its customers to realize these benefits”.
However, Mr. LIND thinks that the neural network at this stage of development is still not perfect and require human efforts: “ultimately, the results of the analysis of neural networks need to be monitored prior to use in trading”.
We conducted a small survey among private traders and found that the vast majority of very indirectly familiar in its analysis of AI systems. Some noted that they use in their forecasts services based on AI, and Gordon JONES, the founder of the DLT project in South Carolina, said: “We concentrated on the aspect of machine learning of all neural networks and AI systems, where they are intended for the analysis of actions and their effectiveness or ineffectiveness, which ultimately leads to a positive or negative ROI”.
Traders are afraid that they will soon be replaced
Managers of top hedge funds in 2015 earned $1 billion, which creates for companies a greater incentive to reduce costs for employees who earn $500 per hour and replace them with neural networks. And Goldman Sachs went on this way, reducing the number of advisers with 600 traders up to 2. Now the rest of the work is done by robots using machine learning algorithms giving invaluable for the company’s clients. It is easy to imagine that many large companies are rapidly following this example, in order to reduce their costs.
In an exclusive interview for Bitnewstoday.ru Amardip SINGH, Professor of artificial technology in Ethereum and analyst at Nasdaq, which has been developing a strategy for the implementation of AI in the economy, said that the traders know that their place will soon take AI and desperately resist the impending change.“Economic analysis is a phenomenon that traders look more broadly: various events, statements from governments that affect the price of what they analyze. In trading there is something called technical analysis. It is mostly very complex patterns which are superimposed on a moving graph, on which are based the forecasts. Technical analysis is also based on human emotions, on how the trader interpreterpath previous market fluctuations. That’s why the AI is incredibly powerful man.”
As being trained and employed neural networks
Thanks to machine learning neural networks allow much more efficient to build non-linear dependence in comparison with the linear methods of statistics such as linear regression, autoregression, linear discriminant analysis. Any analyst who uses technical analysis will make a successful prediction on the basis of preliminary work AI systems. A nonlinear mapping and data visualization neural networks in the space of a smaller number of nonlinear principal component optimizes its processing.
The main advantage of the neural networks before the previously existing algorithms on the basis of the specified parameters is their ability to mimic the behavior of the economy based on social implications.
However, not all experts agree that neural networks can be given full autonomy in decision-making on the basis of their forecasts. So, Bibin BABU, co-founder of the DLT platform in India, believes that the network is not very successful in predicting the digital currency. “The problem is that most neural networks can only access multiple data types inside of the market, not outside. Thus, the forces of crowd psychology, consumer confidence, the effect is good or bad headlines, political or regulatory decisions, the size of the cryptocurrency network, the number of users and merchant acceptance, not fully visible to him. Possible internal market data by themselves are not enough to make any long term predictions or to anticipate short-term price fluctuations”.
Mark LIND believes the ability to use neural networks in trading an art, as a determining factor for the qualitative analysis is the ability to determine the context and the amount of data and their relevance to the type of required results. “It depends on whether you need a search or sorting, and then to determine the context and quality of data. Actually it’s more complicated but at a high level of analysis that brings clarity. You are using mathematical functions, called neurons, which take some numbers as input, and then use the formula for linear combinations to be multiplied by the appropriate weight and then sum it. In fact, the network transforms the data, until you can classify them as output. Do interconnected neurons which generate output, and to use them interactively as input to other neurons, and therefore the entire network. Then the loss function is used to determine how good a neural network to solve a task or problem. Initially, it will not give strong results but over time its use will provide greater accuracy”.
In the next article we will discuss how forecasts are done by our experts regarding the future of the market using neural networks: will the AI manage the market by yourself, or still the final decision will remain with the traders. About this and other concern to all the questions, read the following article.