Constructive Cost Model

Abstract— Software cost estimation is an important phase in software development. It predicts the amount of effort and development time required to build a software system. It is one of the most critical tasks and an accurate estimate provides a strong base to the development procedure. In this paper, the most widely used software cost estimation model, the Constructive Cost Model (COCOMO) is discussed. The model is implemented with the help of artificial neural networks and trained using the perceptron learning algorithm. The COCOMO dataset is used to train and to test the network. The test results from the trained neural network are compared with that of the COCOMO model. The aim of our research is to enhance the estimation accuracy of the

Early Design Model.
2. Post-Architecture Model.
Early Design Model: This model is used to make irregular estimates of a project 's cost and duration before it is entire architecture is not resolved. It uses a small set of new Cost Drivers, and new estimating equations.
Post-Architecture Model: The Post-Architecture model coating the actual development and maintenance of a software product.
Artificial Neural Network is old in effort estimation due to its capacity to learn from previous data. It is also able to model complex connection between the dependent (effort) and independent variables (cost drivers). In addition, it has the ability to derive from the training data set thus enabling it to produce acceptable result for previously invisible data. The goal of the Neural Network is to model the relationship between the input and output from the historic data so that it can be used produce the good estimate for the future projects. Neural Network is compared to regression models and sophisticated Neural Network is better than regression method for estimating effort [6].
III. NEURAL NETWORKS IN PREDICTION
BACK

The network associated with back-propagation learning algorithm is termed as back propagation network. While training a network a set of input-output combination is provided the algorithm provides a procedure for changing the weight in BPN that helps to classify the input output combination correctly.
The aim of the neural network is to train the network to achieve a balance between the net’s capacity to respond and its understanding to give reasonable responses to the input that is similar but not identical to the one that is used in training. Back propagation algorithm modify from the other algorithm by the method of weight calculation during learning. The defect of Back propagation algorithm is that if the hidden layer increases the network become too complex.
IV.DATASET DESCRIPTION
CocomoII The COCOMO Dataset not new in the analysis and acceptance of the model is achieving from the historic projects of NASA. One set of dataset response of 63 projects and other has 93 projects. The datasets is of COCOMO II format. In our measures 93 projects are used for training and 63 projects are used for testing .Number of effort adjustment factor is increases by 5, now it becomes 22 as shown in table 1.
Table 1: Cocomo