What is the best way to start: to put all data I have, or to start with a portion of the data to start building a model?
You can try to process this data in the Eidos system:
https://www.researchgate.net/project/INTELLIGENT-SCALABLE-OPEN-INTERACTIVE-ONLINE-ENVIRONMENT-FOR-TEACHING-AND-RESEARCHING-ON-THE-BASIS-OF-ASC-ANALYSIS-AND-EIDOS-SYSTEM
https://www.researchgate.net/profile/Eugene_Lutsenko/publications
I have never used Statistica 7 software before. So If you want to build neural networks without using any software, here is what I will recommend you to do:
First of all, you need to declare clearly what is your problem? Supervised Learning or Unsupervised Learning, Regression or Classification.
Then, you need a framework to build your neural network (suggest: Python - Keras, Tensorflow, Keras, Pytorch..., Matlab...)
Then, you need to define your features (input of network) and labels (output of network).
Then, If your data is too big (such as > GB, TB data...), you should use 1 portion of your data to test your model first. If your data is enough (such as MB) you should use all of data.
Your portion of data or all of data should split into 3 set, training set, validating set and testing set. Usually, your model has hyper-parameters. Your training set for training model with set of parameters, your validation set for choosing the best set of parameters, and the testing set for test the performance of your model.
Hi,maybe you need to spend hours to study a basic development tool of neural network firstly,pytorch for example,and if you like,YouTube should be a good place for you；
Then follow some simple examples about how to build a deep neural network，pay attention to the structure of the network；
Finally，try to load your data into the example you learn（modify the example or the format of your data if necessary）.
If you want to study more about Deep Learning，please read the relevant literature carefully.
- Badges
- Science topic