TEA Business College: Innovating Investment Analysis

TEA Business College Patents

Intermarket Analysis

Patent Number: US 8,442,891 B2 Patent Date: May 14, 2013

The present invention relates to a method and system for performing intermarket analysis using neural networks. The invention provides a proprietary method and process for selecting relevant markets with the highest correlation in training the neural network from a vast array of available global financial markets, resulting in highly accurate market predictions for each "major" market. The selection process involves identifying "key" intermarkets, "general" intermarkets, and "predictive" intermarkets associated with each "major" market from an available market pool.

The market data for each key intermarket, general intermarket, and predictive intermarket can then be processed to train the neural network, so that when the neural network processes the input data, the neural network generates as accurate output data as possible for each primary market. After training the neural network, all relevant market data for each primary market can be processed through the neural network to predict future market data for each primary market, and then forecast technical indicators can be derived from the predicted future market data for traders to use in making trading decisions.

Calculate forecasted technical indicators.

Patent Number: US 8,560,420 Patent Date: October 15, 2013

The present invention relates to a method and system for calculating forecasted technical analysis indicators. The premise behind technical analysis is that all factors influencing a particular market at any given time are already reflected in the price of that market. Technology-oriented traders employ various computational methods, focusing on the use of various technical studies and indicators to analyze market behavior.

Some common technical indicators include trend indicators, momentum indicators, and volatility indicators. Many technical indicators, such as moving averages, attempt to filter out short-term price fluctuations in order to observe underlying trends. One side effect of doing so is that technical indicators often lag behind the market. Such indicators are referred to as trend-following or lagging indicators. This lagging effect leads to traders reacting to market changes later, resulting in missed profit opportunities and increased risk of losses.

The present invention overcomes this lagging effect by developing methods, systems, and devices for calculating forecasted (leading) technical indicators that do not lag behind the market, based on a combination of historical and forecasted data derived from neural networks applied to intermarket data related to each specific primary market.

In one aspect of the present invention, a method is used to combine forecasted data with conventional technical indicator information using an algorithm to obtain forecasted technical indicators that can guide market behavior, thereby overcoming limitations previously associated with lagging effects.

In the early years, much of Mr. Mendelson's research was done through extensive experimentation and computer power, and was very labor-intensive. Recently, with significant financial investment in the most advanced computer servers and the development of a highly complex proprietary internal research software training platform, the Predictive Technology Group has achieved the automation of much of the mathematical processes associated with executing necessary steps for intermarket analysis and generating forecasted technical indicators, making extremely accurate short-term market predictions possible. This automation is achieved by utilizing servers as intelligent robots.

In the future, the ultimate application neural networks used for training and selecting future trading software programs developed by Mr. Mendelsohn and his research team will require minimal human intervention or judgment decisions. Since Mr. Mendelsohn first began applying neural network pattern recognition to global intermarket data, this complex research methodology has continuously evolved and improved over the past 30 years, costing millions of dollars in research and development. It was the subject of two high-tech patent applications by Mr. Mendelsohn, submitted to the United States Patent and Trademark Office on December 7, 2009, where he first revealed the workings of his research techniques.