Tag Archives: neural network

Release of SMFT-1 Updated Version

Addaptron Software is pleased to announce the release of a new 2013.05 version of Stock Market Forecast Tools SMFT-1. It is an integrated system that includes three major programs: the most popular software program SMAP-3 for stock market cycles analysis and forecast, NNSTP – Neural Network Stock Trend Predictor, and FTA-2 – a modified version of InvAn-4 that is a comprehensive tool used by serious investors for years. The new version includes several improvements, such as, models optimization, ability to read more different input file formats, and optional feature to enable a free Downloader.

Release of SMFT-1 Updated Version

SMFT-1 consists of the tools that employ fundamental ratios rating model, technical analysis, chart pattern analysis, Elliott Wave theory, cycle analysis, candlesticks model, trend lines analysis, regression models, etc. The calculations are empowered by Neural Network. The implemented methods are statistically proven and widely used.

Most methods are provided with back-test calculation to estimate the accuracy of forecast within the recent performance period. The back-testing computations also may play an important role if more than one method is used. It allows estimating a weight of each method in a composed result; the weights that are proportional to the ability of the methods to predict the price.

Advanced Software System for Professional and Institutional Traders

Advanced Software System for Professional and Institutional Traders

After years of intensive development Addaptron Software has released Trading Decision Support System TraDeSS-1, advanced computer program for institutional traders and investors. It is a comprehensive and effective software to help finding the best trading opportunities, maximizing profitability using several predictive models with back-testing features, and optimizing algorithms by running simulations.

TraDeSS-1 is equipped with an advanced forecasting state-of-the-art system. The predicting can be done using nine forecast methods of different nature. Only one method, a combination of a few ones, or all together can be used. Each method is provided with back-test calculation to estimate the accuracy of forecast within the recent performance period. The back-testing computations play an important role if more than one method is selected. It allows assigning a weight to each method in a composed result; the weights are proportional to the ability of the methods to predict the price.

The comparative analysis of simulations shows that systems based on predicted entry-exit signals generate a better profit in around 70% cases than random-entry trading systems. As well as, it should be noted that a multi-model forecast provides a significant improvement over the best individual forecast. It can be explained by the existence of many different independent factors contributing to the error in each forecast which is normally distributed around an actual value.

Since sometimes predictions can fail, to preserve a capital amount in a volatile market, the software enables simulating different risk management approaches. Depending on the character of particular trading shares and the current market conditions some ideas can work better than others. To optimize the strategy in a particular case, the software enables testing different algorithm configurations and finding automatically the best ones. It is especially important for exit points to minimize losses (and ultimately maximize an overall profit). All optimizations can be done automatically by scanning 64 possible logical combinations and adjusting numerical parameters.

TraDeSS-1 has a functionality that allows estimating a hypothetical maximum possible profit in case of 100% accurate forecast. Although an actual forecast cannot be so accurate, this feature combined with comparative analysis enables discovering the best trading opportunities among different types of financial instruments. Calculating maximum theoretical return allows finding optimal buy and sell signals. Also it helps estimating a reasonable amount of initial investment at given transaction fee. Users can choose to re-invest each time a new or the same amount and see the difference in results.

The software has forward testing and assets management features. It allows monitoring the simulated or actual completed transactions, reflecting total trading activity, and evaluating the success of trading in overall. It enables working with many separate data files that is convenient in case of managing multiple assets and keeping the archives of older activities.

TraDeSS-1 has also a few independent tools, such as, technical indicators predictor, cycle analysis forecast, Neural Network (NN) forecast, fundamental 3-month rating model, etc. The detailed description is presented in User’s Manual (accessible from menu Help after downloading and installing the software). To download the software, the registration is required: http://www.addaptron.com/m/registr.php

Candlestick Patterns Decoded by Neural Network

Candlestick patterns can signal a trend continuation or reversal. Many traders and investors know how to benefit from candlesticks analysis. As a rule, a simple chart analysis is used that requires analysts to remember and correctly interpret, at least, typical candlestick patterns. However, the number of possible different candlestick shapes and their combinations in a row can be huge. Besides, some patterns interpretations might be considered as contradictory or doubtful.

Fortunately, statistical methods combined with computer power can be a good solution to make the candlestick patterns recognition works less time-consuming and more effective. These days, Neural Network (NN) can help to automate a candlestick patterns recognition task. NN should be properly trained in order to be able to predict the next candlestick parameters with the highest probability. One of the obvious problems of implementing a candlestick pattern NN predicting system is a formalization of inputs.

Numerous tests show many possibilities of improving NN candlestick patterns recognition systems. For example, output result can be composed from selected optimized calculations based on different historical periods. Also there are many different ways to formalize the shapes and relative positions of candlesticks. As well as, as back- and forward-testing show, it is reasonable to re-train NN for each particular type of shares and latest historical periods to make the forecast more accurate.

Optimal Solution. There is an automated tool FTA-2 (free use of fully-functional version for one month). It has module which enables using Neural Network to recognize typical candlestick patterns and predict future prices (open, high, low, close). This module predicts only one next candlestick but the prediction can be successfully used for different widths of candlestick, i.e., the number of trading days in one candlestick. The module has been enhanced to calculate result that is composed from different historical periods that allows making the forecast more accurate. Also it can perform comparative forecast analysis for many symbols.

Candlestick Patterns Decoded by Neural Network

Useful resources:

  • Candlestick basics – major signals
  • Neural Network basics – introduction
  • The computer program which enables using Neural Network to recognize typical candles patterns and predict future prices – software FTA-2

Using Parabolic SAR with Neural Network for Predicting

Parabolic SAR (SAR stands for Stop-And-Reverse) is a trend-following indicator that has been used by many traders for decades. Its major application is in trading systems to define a trailing stop, i.e., to protect profit when a price trend changes. The term “parabolic” appeared to characterize the indicator parabola shape that is due to using an accelerating factor in the formula. SAR is especially effective in a trending market. To make it more effective in a sideways market, it is often used in conjunction with other indicators.

SAR indicator gives a strong signal when a price trend is about to reverse, therefore, this indicator can be used for prediction. To compare the predictive ability of SAR with other indicators, it has been implemented into the technical analysis module of Fundamental-Technical Analyzer FTA-2. SAR calculations have been used to collect statistics based on the forecast simulations for major indexes and ETFs during August-October 2011 period. As a result, SAR’s position was mostly in “top ten” indicators list.

Using Parabolic SAR with Neural Network for Predicting

The research and presented chart are made by Fundamental-Technical Analyzer FTA-2, one of the software modules that enables composing Neural Network forecasts of many indicators with weights accordingly to each indicator’s predictive ability.

Omitting logical rules for accelerating factor and reversal conditions, a recurring core formula for SAR is the following:

SAR (current point) = AF * [EP – SAR (previous point)] + SAR (previous point)

where:
AF – Acceleration Factor (normally starts from 0.02 and increases by 0.02 if each next point reaches a new extreme, saturates until 0.2);
EP – Extreme Point (lowest low or highest high).

To summarize, Parabolic SAR can be enriched by combining it with Neural Network and successfully used for predicting stock market prices. Combing it with Neural Network allows extracting more statistically stable patterns and, therefore, providing a better accuracy in the forecast. As simulations showed, improved results can be achieved if SAR is transformed into more sensitive indicator by subtracting it from close price (it indicates the degree of SAR and price convergence).

Divergence Indicators Demonstrate Better Predictive Abilities

Technical analysts use many different indicators. Not all indicators are equally good. Some of them have better predictive abilities than others at given conditions. Statistical results of the research for stocks and indexes during the last several months have showed the way to improve some regular indicators. In short, if an indicator is trend-differentially coupled with price – it demonstrates better predictive abilities than a pure indicator.

Calculating Divergence IndicatorAmong studied indicators are Relative Strength Index (RSI) and Moving Average Convergence/Divergence (MACD) indicators. They have been transformed to a slope of line and differentially-coupled with a price line slope. Indicators and price transformation to line slopes has been performed using Least Squares Linear Regression within a sliding 10-day period (moving window). A comparative analysis of indicators’ predictive abilities showed that these two coupled ones are better than around 90% of all other (57) the most popular indicators.

The chart below shows an example of such forecast:

The research results presented here have been calculated using a feature of Technical Analyzer TA-1 (TA). Except analyzing chart with indicators and historical data, TA enables to perform a 10-day forecast using Artificial Neural Network. The calculations can be done on the basis of one selected technical indicator or all available. If all indicators used, TA decides how much weight should be assigned for each indicator’s forecast in a composite result by using back-testing for particular market conditions and a specific stock. Each weight is proportional to a predictive ability of a corresponding technical indicator.

Neural Network vs. Cycle Analysis to Predict the Stock Market

The purpose of this research was to compare two statistical methods: one that based on Cycle Analysis, another – on a simple Neural Network. Price and volume data were used to train this particular Neural Network. These statistical forecasts were built using historical data of S&P-500 index for six months (from June 2009 to January 2010).

The charts below shows how actual 5-day performance (yellow line) differ from predicted performances by these two methods. The top half is the comparison of Neural Network prediction, bottom half – Cycle Analysis. Green bars mean buy signals, red – sell*.

Neural Network vs. Cycle Analysis to Predict the Stock Market

Three major conclusions for this particular historical period:

  1. Cycle Analysis prediction gives signals too early, Neural Network prediction – too late.
  2. In average, the prediction by Cycle Analysis showed slightly better accuracy than the one by Neural Network.
  3. It seems logical to combine these two methods to improve the accuracy.

* ) The calculations have been performed by an integrated experimental system that combined two applications: Cycle Analysis predictor SMAP-3 and Neural Network predictor NNSTP-2