I have the historical data about the stock prices of the companies and can easily calculate the historical correlation between any two of them.
Given this information, how should I go about finding the position sizes that will provide me the maximum diversification?
Example: Consider three companies, out of which two of them are real estate companies and the third one is an automobile company. The correlation between the two real estate companies is higher[as found from the data]. My intuition tells me that I should give less than (1/3) weight to each of the real estate companies, in order to achieve higher diversification.