Neural net backpropogation as per EasyNN-plus

Hi all. I'm not sure if anyone has heard of EasyNN-plus (EasyNN-plus) but I'm trying to build something similar in Matlab. According to the FAQ EasyNN-plus uses a backpropagation neural net and a transfer function of logistic function. That is 1.0 / (1.0 + e (-net input)) . I'm trying to design something in Matlab that would do the same. It doesn't seem to be possible in the Deep Network Designer as this doesn't allow data input of the type used in the Regression Learner for example 9 inputs with one output i.e. not graphical input or time series. I may have completly missed what is already possible so if someone could point me in the right direction it would be most appreciated.

SPG

Answer 1

[Stephen Gray]

Looking at this again, I think I need to be using feedforwardnet. The problem is fitting my data to the lines

net = feedforwardnet(10);
net = train(net,x,t);

The input data is 9 fields * 1000 lines and the target data is one field * 1000 lines i.e. 8 inputs and 1 output. The problem is getting this data into the train function as I'm not quite sure what :-

The cell array format is more general, and more convenient for networks with multiple inputs and outputs,
allowing sequences of inputs to be presented. Each element X{i,ts} is an Ri-by-Q matrix, where Ri = net.inputs{i}.size.

means.

SPG

Answer 2

[Shujaat]

https://techpassion.co.uk/how-does-a-smart-tv-work-read-complete-details/

I am writing to inquire about the recent integration of new technologies into MATLAB. Specifically, I am interested in learning more about any advancements or updates related to machine learning, AI, or real-time data processing that have been incorporated into MATLAB’s toolboxes. Additionally, I would appreciate any information on upcoming features or collaborations with emerging technologies. Could you please provide some details or direct me to relevant resources?

Answer 3

[Shujaat]

.Thank you for reaching out. We are excited to inform you about the latest advancements in MATLAB. Recent updates include enhanced capabilities in machine learning and AI, with new functions in the Deep Learning Toolbox and improved integration with frameworks like TensorFlow and PyTorch. Additionally, we have introduced real-time data processing features in the Data Acquisition Toolbox, allowing seamless integration with IoT devices and sensors. You can explore these technologies through our official documentation and webinars available on the MATLAB Central website. Feel free to reach out with any further questions.