Inference of Media Bias and Content Quality Using Natural-Language Processing