Author-Specific Linguistic Patterns Unveiled: A Deep Learning Study on Word Class Distributions