Estimating the Strength of Authorship Evidence with a Deep-Learning-Based Approach