The Research Of POPs Software Estimation Based On Multiple Linear Regression Model

Software Estimation is based on the content of development, development tools,developers and other factors, to forecast the time and workload which is taken by thesoftware requirement, programming design, coding, testing and the whole process.Software Estimation is the basis for software project designing, however, the risk ofsoftware estimation is one of the biggest risks, so Software Estimation has been thefocus of the research in the area of software engineering.The emergence of object oriented software development methodology, makesthe traditional software estimation methods, such as Lines of Code, Function pointmethod, not satisfy the requirements of the characteristics of object oriented software;however, the majority of existing object oriented software estimation methods such as3D Function Point method, Use Case Point method, can only carry out a unilateralmeasure, also can not meet the requirement of software estimations. Predictive ObjectPoints (referred to as POPs) method is a comprehensive software metrics estimationmethod for object oriented features; tend to object oriented software designing. Itconducts a comprehensive measure from the function points of software, theinformation between object exchanges and the impact of succession to reuse,forecasting the scale of software. The core of POPs method is the Weighted Methodsper Class (referred to as WMC). Traditionally, we use chart to count the weight ofeach method and the proportion of every kind of method ,for estimating WMC. Thismethod needs large quantity of historical data, and the software type and industry ofthe historical data covered are not accurate, so the accuracy of the calculation islow [1].This paper researches Software Estimation methods and Mathematics Modelingmethods of software application data, in particular, POPs method and Multiple LinearRegression Model. To resolve the problem of low accuracy of estimation in POPsmethod, proposes POPs software estimation method based on Multiple Linear Regression Model. The main research work we do is as follows: First, we proposePOPs Software Estimation method based on Multiple Linear Regression Model andgive algorithms, introduce calibration. Second, for shipping logistics managementsystem, design and realize the software estimation system of shipping and logisticsmanagement. At last, we apply POPs method to specific area Shipping LogisticsManagement; make a comparative analysis for the experimental data and results,improving the accuracy of the estimation.