Which tool should I use for Software effort estimation---nftool, nprtool or ntstool?

I am confused when to use which tool.

I WANT to predict (estimate) software effort based on some other inputs.

Effort is always a continuous number while other inputs may be continuous or categorial.

Please help me which tool is best suited to my problem.

Answer 1

[Jayanti]

Hi Sushma,

Based on your requirement "nftool" in MATLAB seems to be a right choice. "nftool" is specifically designed for tasks, where the goal is to estimate or predict a continuous numeric value, such as effort .

"nprtool" is for classification tasks, which is suitable when predicting discrete categories not suitable for continuous effort values. "ntstool" is used for time-series data. Unless your data is explicitly time-dependent, this tool is not ideal.

Since inputs can be continuous or categorical, you can encode them using one-hot encoding before feeding to the network.

I am also attaching the documentation link of "nftool", "nprtool" and "ntstool" below for your reference:

https://www.mathworks.com/help/deeplearning/ref/nftool.html

https://www.mathworks.com/help/deeplearning/ref/nprtool.html

https://www.mathworks.com/help/deeplearning/ref/ntstool.html