Research on Mathematical Statistics Method Based on Computer Big Data

The software measurement model under computer big data is based on the measurement value of the software engineering project (such as the size of the project team) for the future software engineering to predict the target software measurement value, such as the software development workload and error rate statistics. Obviously, building such a software measurement model requires the use of data samples from similar projects in the past. However, there are often missing data in these data samples. The determination of which measurement variables to choose as independent variables in regression modeling is likely to be based on intuition or empirical assumptions. After the model is established, empirical hypotheses are rarely tested after the fact, so it is easy to produce redundant measurement variables and increase unnecessary complexity. For this reason, the article mainly considers the processing methods of discrete variables. Introduce a more common method of dummy variables. Finally, the article uses the method of case analysis and uses computer big data R, SPSS, Java and other languages and programs to complete the simplification of the software measurement data model under the absence of data.