Study On Cost Estimation Model For INMS Project Based On Regression Analysis

With the continued development of the global communications technology, themarket competition is becoming increasingly fierce. And the communications industryhas entered the era of little profit, facing tough cost challenges. Because of the specialnature of INMS project, the implementation schedule of the project is very tight and thetasks are heavy, obtaining little time to estimate project cost. Spending more time onproject cost estimation will lead to lose opportunities and place. Developing a fast andreliable project cost estimation methods is of great significance to the company under thenew situation. Since the methods of regression analysis on domestic telecom projects isextremely rare, this paper is intended to be the introduction of this cost estimationmethods and philosophy for domestic telecom projects, and providing useful lessons forthe company.In this paper, the theory on the basis of in-depth research project cost estimation andstatistical analysis, a cost estimation model is built by the practical experience of theestablishment, inspection, improve and practice as a whole process. In the process of themodel building, the paper solves the three problems: key complexity factor of selection,the statistical significance and the model empirical significance. In the study on the keycomplexity factor selection, the paper leads to a detailed analysis on INMS projectconstitution, in-depth and detailed understanding of its required workload on the keyfunction and development. Comparing the importance degree of the various parts, thepaper finally quantify the key complexity factors which impact the project cost heavily,combined with advises given by the experts. In the study on validating the model of thestatistical significance, the model is established on the basis of mathematical methods instatistics, and amended by the method of the statistical test. In line with the statisticalsignificance, the model is inspected by the actual historical data afterwards, bycomparison with the results of the original project evaluation methods, detecting modelperformance in actual use. As a result of the model performance for INMS project inactual use, the parametric model improves the cost estimation performance for INMSproject from different perspectives. Compared with the original estimation method, the improved model on calculating costs saves half time and the error rate is reduced to onethird of the original measure.Although this INMS project cost estimation method has made certain achievements,the research on the history of the project data only relies on statistical linear regressionmethod, for this reason that the model has a few limitations of the predictive capabilityand the use of environment. The model in the future research work, according to thesustainable development of the project cost estimation techniques and methods, should beimproved and be optimized continuously.