## Computerized Trading: Maximizing Day Trading and Overnight ProfitsComplete with explanatory charts, diagrams, and checklists, a detailed discussion of the use of computerized trading systems draws on the expertise of twenty expert traders, and gives step-by-step advice on strategies and reliable data to both beginning and experienced investors. |

### From inside the book

Results 1-3 of 8

Page 296

To address the problems that backpropagation neural networks present, a new

class of neural networks has been developed, called ontogenic statistical

networks. The name comes from the word ontogenesir meaning the history of the

organism. An ontogenic network develops its own topology during training. No

longer does the developer need to decide, a priori, how many layers, what

transfer functions, or how many

network technology ...

To address the problems that backpropagation neural networks present, a new

class of neural networks has been developed, called ontogenic statistical

networks. The name comes from the word ontogenesir meaning the history of the

organism. An ontogenic network develops its own topology during training. No

longer does the developer need to decide, a priori, how many layers, what

transfer functions, or how many

**nodes**. Depending on the specific ontogenicnetwork technology ...

Page 297

The three input

upper right. The intersection of the lines represents the weights. After the initial

training, if the problem is linearly separable, the problem is solved. If there is

residual error, then hidden

error to acceptable levels. A pool of potential hidden

random initial weights. All

training cycles is ...

The three input

**nodes**are shown on the left and the output**nodes**are on theupper right. The intersection of the lines represents the weights. After the initial

training, if the problem is linearly separable, the problem is solved. If there is

residual error, then hidden

**nodes**must be added (one at a time) to reduce theerror to acceptable levels. A pool of potential hidden

**nodes**is created usingrandom initial weights. All

**nodes**are trained until either the maximum number oftraining cycles is ...

Page 298

Assume that our data is a perfect sine wave and we want to model this using

Cascade Correlation and a sine as the

Correlation cannot do this. Remember that the first step was to do a linear fit and

then model the residual error with the hidden

17.3 represents the linear fit. Once this linear fit is accomplished, the hidden

Assume that our data is a perfect sine wave and we want to model this using

Cascade Correlation and a sine as the

**node's**transfer function. With one hidden**node**, we should be able to exactly model this data. However, CascadeCorrelation cannot do this. Remember that the first step was to do a linear fit and

then model the residual error with the hidden

**nodes**. The straight line in Figure17.3 represents the linear fit. Once this linear fit is accomplished, the hidden

**nodes**must model ...### What people are saying - Write a review

We haven't found any reviews in the usual places.

### Contents

PART l | 1 |

Popular Trading Indicators | 13 |

Combining Trend Analysis with Indicator Readings | 30 |

Copyright | |

22 other sections not shown

### Common terms and phrases

analysis approach backtesting bars beneﬁt Bollinger Bands breakout buy signal chart close congestion contract data mining data vendors datafeed deﬁned develop diversiﬁcation drawdown equity curve evaluation example Exchange exit Exponential Moving Average Figure ﬁlter ﬁnancial ﬁnd ﬁrst ﬁt ﬁve formula frame futures futures contract fuzzy logic genetic algorithms Index input Intermarket intraday investors linear losing trades loss method momentum money management moving average neural networks nodes nontrending Omega Research optimization options oscillator out-of-sample pattern percent period portfolio position predict proﬁt proﬁtable ratio real-time Relative Strength Index risk sell signals signiﬁcant simple moving average speciﬁc Statistical Network Steve Fossett stochastic stop T-bond Table technical analysis technical indicators techniques tick tion TradeStation trading strategy trading system trend trendline uptrend variables volatility volume winning trades zone