Abstract

Uncertainty is the essential characteristics of the data, the analysis of the uncertain data has been concerned in more and more areas. Traditional uncertain data is represented as a range and the probability distribution function over this range, because the uncertain data exists both fuzziness and randomness, so the actual distribution of the uncertain data is difficult to be accurately defined by the traditional probability distribution function. Therefore, an uncertain data cloud modeling process is proposed based on the distribution of cloud droplets in cloud model, and through cloud integration and cloud similarity computing to achieve the classification of uncertain data, the cloud model can reflect the actual distribution of the data, so it will achieve data classification more effectively, our experiments also proved the effectiveness of this method.

Paper Details