Example:
Linear Regression Matrix of financial data:
The Regression Analysis can help us to find the correlation of data. Usually we fit
these test data to a linear or a polynomial function. How good the data fit to this
function is described by the regression coefficient.
Stock data are almost unpredictable. But companies usually are tied to a market. The
market can change
or other markets can have an influence on a market, which is often difficult to
understand. When people's income is reduced by an increase of energy prices or there is a
fear of unemployment, stocks of companies can drop or rise. Also big money movers or
political events can influence these markets. Therefore it is very helpful for investors
to find out, which companies are related, either because they depend on the same market or
they are aim of a strategic money flow.
The regression matrix can help to find the correlation between companies. Such
relations can be temporary or constant over time.
In the next picture, is an example of a regression matrix of some companies of the Dow
Jones Industrial made from data from the year 1998 to 2003..
We created a regression matrix of the stock data of 7 companies of the "DJ" over 5 years. One company showed no correlation . Walmart and United Technology had a regession coefficent of 0.5. When we look at the stock values of the last 3 month , we see, that the stocks of both companies were in parallel except for about one month.Very uncorrelated data of companies like Walmart and Disney are shown in the next picture.
ELB 06/02/2003