Factor models are helpful for simplifying complex data. A new method was developed to estimate these models in high-dimensional datasets, like financial data. By comparing the spectral patterns of the data with a model, the number of factors and correlation structures can be determined. This method is effective in noisy data and captures important market dynamics.