Clara Algorithm in A Distributed System

The development of websites, Facebook websites and social network websites are an extremely fast way of commerce, education, so on. The effective uses of billions of documents on the websites, Facebook websites and social network websites are a very important significant for many commercial applications, many personal applications and many researches in a long time. With these reasons, in this research, we propose a new model using Clara Algorithm with Hadoop Map (M) /Reduce (R) for English document semantic classification in distributed system – a parallel network environment. Our new model can be used in classifying billions of English documents in a short time in a distributed system. We test our new model on our testing data set (including 25,000 English reviews which have 12,500 positive English reviews and 12,500 negative English reviews) and achieved on 60.3% accuracy. Our English training data set has 70,000 English sentences, including 35,000 positive English sentences and 35,000 negative English sentences.