|
Abstract:
|
Interval data deals with numbers, which are not single values but sets of numbers ranging between a maximum and a minimum value. There are usually two situations that carrying out interval data as a result. One is throughout the data collection process as collecting blood pressure. The other is due to the data process as large data set summarized into a manageable size. Because of the special format of interval data, the traditional statistical analysis method may not suit very well. The existing methods are almost all based on the assumption of uniformly distributed data points inside intervals, which is not necessarily true in real data world. For the case that both X and Y are interval data, in order to find out the correlation between X and Y, we proposed a method of resampling as resampling data points based on the assumed distribution of intervals to estimate correlation and build model. For interval data, the information known are mainly the lower and upper bounds of intervals, as well as the distribution of the data points inside intervals. By using resampling method, we are able to break the assumption of uniformly distributed data points.
|
