Neural Networks and FinBrain’s NYSE Index Tickers Prediction Performance – 07/02/2018-07/09/2018

FinBrain Technologies develops high performance and high precision Deep Learning Technologies for analyzing and predicting the price movements of Financial assets. We constantly improve our algorithms and add new features to our services. FinBrain collects and processes vast amounts of Financial Data for every single asset listed under its services. Our Deep Learning models “learn” from these Financial Datasets, extract features and expose hidden patterns, calculate correlations and assign more weights to the most important predictors of the future price movements.

The Artificial Neural Networks model the human brain’s learning process in Mathematical ways. Biological neurons receive the electrical signals from dendrites, the cell body processes these signals and checks whether the signal strength is larger than the threshold value. If the signal strength is larger than the threshold value, the signal is transmitted through the axon and the signal is sent to the other neurons.

Just like the biological neurons, the artificial neurons have many input nodes to receive input data from various sources/features. The artificial neuron assigns weights(w) to each of these inputs(x), then adds a bias value(b) sums the input-weight product values and biases for each input and passes this information(x*w+b) through an activation function.

The neuron assigns weights to each one of the inputs. The input value is multiplied by a weight value, and during the “learning” process of the Neural Network, the neurons are adjusting their weight values to minimize the difference between the target and test values. The inputs multiplied by the weights are summed up and an offset(bias) value is added. The bias value is also adjusted through the “learning” phase. At the beginning of the Neural Network “training”, the weights and biases are randomly initialized and during the learning, these values are gradually shifted to reach the desired result value. The result of the calculations are then fed into an activation function to obtain the output.

The Multi-Layer(Deep) Neural Network structure is shown above. Deep Neural Networks consist of multiple layers having multiple neurons at each, and each neuron in one layer connected to the neurons in the next layer. Using this architecture, different kinds of financial data can be analyzed and the outcomes can be predicted.

FinBrain Technologies creates deep neural network structures that analyze and predict the price movements of more than 6000 Financial Assets including stocks, cryptocurrencies and foreign currencies. The assets are grouped under S&P500, NASDAQ, NYSE, Cryptocurrency and Foreign Exchange Markets, and the financial data and predictions can be obtained by subscribing to FinBrain’s services on https://finbrain.tech/#/pricing.

We keep posting the past prediction performance results on our blog for both markets. FinBrain does not only provide BUY or SELL signals, but also MODELS the future price movements of the assets. We do not draw lines, support or resistance levels or say that the stock might go higher if it breaks the resistance. FinBrain does not look at a few indicators and say that “this might be a sell point”. Instead, FinBrain analyzes how different types of financial data determines the price outcomes of the assets. This is way beyond human capacity to analyze and interpret, as human brain cannot feature in many different aspects at the same time and calculate how these affect the price outcome.

3-Day Change / Expected: 6.12% vs Real: 6.22%

5-Day Change / Expected: 9.62% vs Real: 19.82%

3-Day Change / Expected: 1.48% vs Real: 1.87%

5-Day Change / Expected: 4.68% vs Real: 11.25%

3-Day Change / Expected: 8.78% vs Real: 5.43%

5-Day Change / Expected: 13.93% vs Real: 9.58%

3-Day Change / Expected: 5.52% vs Real: 4.80%

5-Day Change / Expected: 8.38% vs Real: 8.66%

3-Day Change / Expected: 1.86% vs Real: 2.97%

5-Day Change / Expected: 4.39% vs Real: 8.19%

 

3-Day Change / Expected: 5.36% vs Real: 3.00%

5-Day Change / Expected: 7.40% vs Real: 7.84%

3-Day Change / Expected: 2.48% vs Real: 4.03%

5-Day Change / Expected: 4.60% vs Real: 7.17%

 

3-Day Change / Expected: 2.72% vs Real: 4.11%

5-Day Change / Expected: 4.30% vs Real: 6.92%

3-Day Change / Expected: 3.22% vs Real: 4.42%

5-Day Change / Expected: 3.66% vs Real: 6.57%

3-Day Change / Expected: 5.10% vs Real: 3.65%

5-Day Change / Expected: 7.10% vs Real: 6.33%

FinBrain Technologies is one of the pioneers in integrating Deep Neural Networks in Financial Analysis and Prediction. As we always emphasize, AI is changing the world and early adopters have the advantage. Sign up to FinBrain on https://finbrain.tech , subscribe to a package now on https://finbrain.tech/#/pricing  and beat the market.

Please feel free to contact us at [email protected]

FinBrain Technologies, 2018

Leave a Reply

Your email address will not be published. Required fields are marked *