This is a trading item or a component that was created using QuantShare by one of our members.
This item can be downloaded and used by QuantShare Trading Software.
Trading items are of different types. There are data downloaders, trading indicators, trading systems, watchlists, composites/indices...
You can use this item and hundreds of others for free by downloading QuantShare.
Top Reasons Why You Should Use QuantShare:
Works with US and international markets (stock, forex, options, futures, ETF...)
Offers you the tools that will help you become a profitable trader
Allows you to implement any trading ideas
Exchange items and ideas with other QuantShare users
Our support team is very responsive and will answer any of your questions
We will implement any features you suggest
Very low price and much more features than the majority of other trading software
For Free - No Credit Card Required
Coefficient of Determination - R Squared - Time-Series Prediction
In statistics, the coefficient of determination represents the strength of the relationship or the portion of common variation in two time-series or variables. It is a statistical measure of how well the regression line approximates the real values.
The coefficient of determination or RČ is mainly used to analyze how well a variable can predict another one. The returned value gives us the percentage of change of variable X that can be explained by changes in variable Y.
The coefficient of determination is calculated by taking the square of the Pearson correlation (simple linear correlation). A value of 1 or 100% indicates that moves in one time-series can be entirely explained by the second time-series, while a value of 0 or 0% means that that the first time-series cannot be explained by moves in the second time-series. RČ values range from 0 to 1.
Example:
The correlation between the weekly change of the S&P 500 Index (Ticker Symbol: ^GSPC) and the previous week change for the last 50 trading weeks is equal to 0.13.
The coefficient of determination is equal to 0.0169 or 1.69%, which means that weekly changes can explain only 1.69% of next week changes.
Formula Example:
a = r2(perf(close, 1), perf(close[1], 1), 50);
plot(a, "", colorGreen);
Calculate RČ of the current and previous rate of change using the last 50 trading bars.
Plot the coefficient of determination of each trading bar on a chart.
