Mean-variance data collections for portfolio optimization problems

Mean-variance data collections for portfolio optimization problems based on time series of daily stock prices for companies traded on major USA stock exchanges. These data collections can be used for investment portfolio optimization research.From 1912 stocks, we randomly selected two sets of size 1000 (denoted by JKMP1 1000_1 and JKMP1 1000_2), such that their intersection is minimal. Form 1912 stocks, we also randomly selected three disjoint sets of size 500 (denoted by JKMP1 500_1, JKMP1 500_2, and JKMP1 500_3). All stocks from JKMP1 500_1 are also in JKMP1 1000_1, and stocks from JKMP1 500_2 are in JKMP1 1000_2. For each set of stocks, we estimated the correlation matrix and the vector of mean returns, based on the corresponding time series.Each data collection is saved in a text file following the format used by J. E. Beasley in OR Library (http://people.brunel.ac.uk/~mastjjb/jeb/orlib/portinfo.html).A detailed description of the data collections can be found in the README file of the dataset.