In my future works I want to apply Artificial Intelligence (AI) methods instead of traditional Statistical Methods
in #literacy, and #debates, as far as X_time.for.learning is concerned or Z_time to.spread.an.educational.message, 'certain that training several _#stack in an _#ANN with #Mathematical #models #can -> #interresting #data.
Fuzzy logic and set theory are the traditional method, while AI is newer and widely accepted method. It can be applied with ease as there are many tools are available. MATLAB comes with a toolbox i.e. "nprtool" and "nntool".
Thank you for your valuable comments @Rita Kushwaha and @ Emmanuel Léo Riviere
For application ANN, many digital data is necessary (for learning, for testing). If you have these data, it is good tool.
FL is suitable when you have some data and questionnaires where the answers to the questions are fuzzy. And yet, here you need to have an expert to determine the the fuzzy sets, fuzzify the input and output data, and also to draw up fuzzy rules.
Fuzzy logic allows you to change the knowledge of an expert, eg about controlling an airplane into a computer program. You enter this program into a computer that will control the plane automatically without human intervention. The same applies to all robots that have to do something. Is this not a valuable scientific method? However, neural networks are self-learning systems from samples. They are also very valuable.
Dear Prof. Ashim Bora,
Let me ask a counterquestion: When making informed decisions in the grey areas of business, which side of the brain do you think is better in the processing of information? The left or right hemisphere?
Dear Prof. Ashim Bora,
Take a look at this sequence of numbers: 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, ...
Assuming that you have no knowledge about this sequence, in less than 10 seconds, your brain can easily "predict" the next few numbers in the run. See how effective your biological neural networks are.
However, if you are asked to find the number at the 100th term in 10 seconds (even with a programmable calculator), you probably fail do so unless you know the generalization of the series. Hence, you'd ask a Math Expert about the "rules" of computing the series.
#RNN#should'nt#they#be#trained#in#precontrainted# A000045 or other sequel, howsoever, all #ia.sequence traine their #stak the same #@try.circle.number, with #datascientist, to normalize #training #nummer? Is there #benchmark for #IA.convergence?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence