Iâ€™m busy working on my S&P500 Volatility Model tonight and Iâ€™m slowly inching toward a better correlation. Despite my headway, it still needs a lot of work to be statistically significant. Right now its close to 75% correlated but I still havenâ€™t figured out how I want to rank the volatility events.

I did create a nice image of how the data looks after its modeled. The x-axis is the rate of price change in the S&500 closing price and the y-axis is the volatility event (1 being severe and 0 being safe). If you squint at the image you can see a possible â€œUâ€ shape outline, which means that this model could be explained by a second order function (parabola).

The next image, a chart from my Excel spreadsheet, shows the parabolic function of the modeled data.

Well, back to work!

[tags]S&P500, Volatility, Models, NeuralNet, Timing, Quantitative, Finance[/tags]