A data difference analysis method and system based on probability density estimation

Abstract

The invention discloses a data difference analysis method and system based on probability density estimation, belonging to the field of data analysis. The method is to establish a data set first, and the data in the data set have changed; then use the probability density estimation method to estimate the joint probability of the data before and after the change; select the optimal window width according to the maximum likelihood method, for different window widths, Take any point in the data set each time, use the remaining points in the data set to construct the joint probability distribution, calculate the joint probability density value of the arbitrary point on the joint probability distribution, and obtain the joint probability density value of multiple joint probability density values The product is the likelihood value, and the window width with the largest likelihood value is the optimal window width; according to the optimal window width, the joint probability density distribution of the data before and after the change is obtained by the probability density estimation method, and the difference of the data is analyzed. This method can obtain the significance degree of each data without being limited by the data distribution, and is used to find significantly changed data.

Description

technical field [0001] The present invention relates to the field of data analysis, and more specifically, to a data difference analysis method and system based on probability density estimation. Background technique [0002] Data that changes significantly are often critical. For example, through protein mass spectrometry, we can obtain the expression levels of each protein in the experimental group and the control group, and proteins with significant differences in expression may play a key regulatory role in this process. People often look for differential proteins based on the multiple of difference, and think that the protein with a larger multiple of change is more significantly different. However, in most cases, this assumption cannot be established. For example, if 1 becomes 2 and 10 becomes 20, the change is 2 times, but it does not mean that the difference is the same in significance. Another example is the amino acid mutation that affects the protein modificatio...

Images