Data Science or Rule Based System?

It can be noticed that both approaches are being used by the participants. Somebody has declared Ideal Data Scientist does not need Training Set while Data Science follower is exploring features of the sequences. Rule based system would require a PhD style study of the 100k+ systems besides taking a long, long computer time for the expert system to evaluate the 100k+ test sequences. OEIS reports 1 paper per sequence and possibly the writeup of these papers involved months of study each. DS style approach assumes that rules of building such sequences could be learned by algorithm like randomForest. Such strategy requires features to be engineered properly. Can we together develop a list of features? Some features based on rule based study could be: 1) First Term, 2) Second last…


Link to Full Article: Data Science or Rule Based System?