Software Project Estimation by Abran Alain

Author:Abran, Alain
Language: eng
Format: epub
ISBN: 9781118959329
Publisher: Wiley
Published: 2015-04-12T16:00:00+00:00

Figure 6.1 Mean Magnitude of Relative Error – MMRE

Prediction Level of the Model

The prediction level of a model is given by , where k is the number of projects in a specific sample of size n for which MRE ≤ l.

It is not sufficient for a productivity model to perform well with the dataset used to build it, but it also has to perform well with other data (e.g., predictability), and this is very dependent on the number and quality of the data sample in input to these models. To analyze the predictability of a productivity model, the model has to be tested with projects other than the ones that contributed to its design.

For such a predictability analysis, model builders would typically build the models using a large subset of the projects (say 60%), and then test the performance of the model with the subset of hold-out projects (i.e., the 40% not used for building the model). Another method is the so-called leave-one-out strategy, where the models are built with n−1 data points, and tested with one left out. This is done iteratively with a different data point left out at each iteration.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.