Contextuall's  Instablog

Many people believe that we live in a world designed for-and increasingly controlled by-algorithms.

That's where Contextuall comes in.

It is our goal to create a massive inventory of algorithms that can be used to predict a wide range of variables-stock market prices, elections, geopolitical conflicts, movie sales, economic data and more.

We've just launched our website, offering free access to backtested performance analysis on each algorithm. Contextuall also offers free financial research reports and market data.

Today we're pleased to announce the launch of 5 new proprietary algorithms to predict the direction of Sirius XM Radio (SIRI).

During an earlier report we noticed that Radio Broadcasting stocks were the most shorted industry in our coverage universe. We wanted to take a closer look, and figure out if the excessive pessimism is misplaced.

We decided to focus our attentions on Sirius XM (SIRI), using the stock as a proxy for the overall industry.

We created 5 algorithms that now go into testing (links listed below). Every day we'll monitor the performance of these algorithms, keeping an eye on their stability as market conditions change.

PLEASE NOTE: All these projections are based on data as of close on Friday, July 27. All these models will be judged on their accuracy from that date.

If you'd like to know more about Contextuall's services, or if you have a specific stock you'd like to model with our system, contact us.

We've also included links to recent Contextuall data feeds at the bottom of this page.

…..

Instance Based Algorithm to Predict Direction of Sirius XM (SIRI)

Data Set Used to Create Model (In Sample): 728 observations between 4/8/2008 and 2/24/2011
Data Set Used to Test Model (Out Sample): 359 observations between 2/25/2011 and 7/13/2012
Overall Accuracy During Out of Sample Test: 63.04%
Up Signal Accuracy in Out of Sample Test: 61.75%
Down Signal Accuracy in Out of Sample Test: 65.15%
Projected Direction of SIRI Over Next 10 Sessions: Down

Classification via Regression Algorithm to Predict Direction of Sirius XM (SIRI)

Data Set Used to Create Model (In Sample): 870 observations between 4/8/2008 and 9/16/2011
Data Set Used to Test Model (Out Sample): 217 observations between 9/19/2011 and 7/13/2012
Overall Accuracy During Out of Sample Test: 62.32%
Up Signal Accuracy in Out of Sample Test: 64.96%
Down Signal Accuracy in Out of Sample Test: 58.89%
Projected Direction of SIRI Over Next 10 Sessions: Down

Entropy-Based Distance Model to Predict Direction of Sirius XM (SIRI)

Data Set Used to Create Model (In Sample): 870 observations between 4/8/2008 and 9/16/2011
Data Set Used to Test Model (Out Sample): 217 observations between 9/19/2011 and 7/13/2012
Overall Accuracy During Out of Sample Test: 64.73%
Up Signal Accuracy in Out of Sample Test: 70.83%
Down Signal Accuracy in Out of Sample Test: 59.46%
Projected Direction of SIRI Over Next 10 Sessions: Up

Stochastic Gradient Descent Algorithm to Predict Direction of Sirius XM (SIRI)

Data Set Used to Create Model (In Sample): 870 observations between 4/8/2008 and 9/16/2011
Data Set Used to Test Model (Out Sample): 217 observations between 9/19/2011 and 7/13/2012
Overall Accuracy During Out of Sample Test: 59.42%
Up Signal Accuracy in Out of Sample Test: 60.74%
Down Signal Accuracy in Out of Sample Test: 56.94%
Projected Direction of SIRI Over Next 10 Sessions: Up

Ensemble Learning Model to Predict Direction of Sirius XM (SIRI)

Data Set Used to Create Model (In Sample): 870 observations between 4/8/2008 and 9/16/2011
Data Set Used to Test Model (Out Sample): 217 observations between 9/19/2011 and 7/13/2012
Overall Accuracy During Out of Sample Test: 66.18%
Up Signal Accuracy in Out of Sample Test: 67.48%
Down Signal Accuracy in Out of Sample Test: 64.29%
Projected Direction of SIRI Over Next 10 Sessions: Up

…..

Other research released by Contextuall over the last 24hrs…

….

Momentum

Bullish Momentum: Industries With Strongest Upside Momentum

Bearish Mode: The Worst Performing Industries (Moving Average Analysis)

Industries in Focus: Best and Worst Industries (Long-Term Momentum)

Medium-Term Focus: The Best and Worst Industries By 50-Day Moving Average Trends

Short-Term Focus: The Best and Worst Industries By 20-Day Moving Average Trends

RSI(14) Momentum: Oversold and Overbought Industries

News Summary

Today's Most Commented News Stories: July 31, 2012

Short Squeezes

Using The Short Ratio to Identify The 10 Most and Least Shorted Industries

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.