Automatic Term Extraction Using Log-Likelihood Based Comparison with General Reference Corpus