This dataset accompanies a research paper on leveraging Machine Learning (ML) techniques for regression to predict the optimum DC bias in direct current in optical orthogonal frequency division multiplexing (DCO-OFDM).  The dataset comprises a set of features to facilitate the prediction of the required DC bias to mitigate the impact of clipping distortion at the transmitter. MATLAB software was utilized for modelling the DCO-OFDM transmission and generating the research dataset.