Feature Combination for Measuring Sentence Similarity