The authors consider enhancing the performance of the traditional linear partial-response (PR) equalizer by using noise prediction techniques. The resulting noise-predictive PR equalizer consists of a forward linear PR equalizer followed by a linear predictor to whiten the noise and the residual distortion at the equalizer output. Assuming correct decisions and applying results from prediction theory, it is shown that the minimum mean-square error of a noise-predictive PR equalizer equals that of a decision-feedback equalizer which employs no PR shaping. This result shows that the performance of a receiver with a PR shaping equalizer can always be improved by attaching a noise-whitening predictor. Simulation results are presented for channels typically encountered in wire transmission and magnetic recording, i.e., channels with a spectral null at DC and strong high-frequency attenuation. The results show a substantial performance improvement when a noise-predictive PR equalizer is used.< >