Recognizing Humor Without Recognizing Meaning

摘要: We present a machine learning approach for classifying sentences as one-liner jokes or normal sentences. use no deep analysis of the meaning to try see if it is humorous, instead we rely on combination simple features these are enough detect humor. Features such word overlap with other jokes, presence words common in ambiguity and idioms turn out be useful. When training testing equal amounts from British National Corpus, classification accuracy 85% achieved.