Cost Estimation

Project Aim

Using multi-layer neural networks to estimate cost of construction of a building based on multiple parameters.

Members

• Shishir S Volety
• Sathvika
• Saurabh
• Shashank
• Shreehari

Work Done

We implemented gradient descent using the sigmid function and choosing a suitable loss function iterated multiple times till the error between the predicted cost and known cost has a minimal error difference between them.

Future Scope

Implement optimization techniques such as hyper-parameter training, regulation and improve the quality of our data set by adding more data points to train on to increase the generalization ability of the network.