I am using a Genetic Programming Concept in my Embedded project... so far .. It works fine.. But I am still confused with some stages...because a lot of logistics operations are there in my project....
hello friend
For self learning system, Bayesian algorithms will give good results.
But as I dont know the exact problem that you are trying to solve, better you try some huerestics methods.
All the best.
Best regards
Dr.Indrajit Mandal
i have seen some research on machine learning using
hidden markov model
specially when you need to do pattern recognition
Learning classifier systems are a well-established, but often overlooked technique as although they are powerful, they have a bit of a learning curve. They combine the global search capabilities of evolutionary computation with the local search capabilities of reinforcement learning. So as well as effective learners, they have the ability to divide-and-conquer the search space automatically into niches optimizing areas of search independently. Training with either online or offline data produces Human-interpretable ‘if then’ rules, which can be adjusted and run in pseudo-real-time on embedded systems.
Dear Bipin
This article gives a good overview of adaptive learning algorithms and their implementation to supplement your genetic algorithm:
http://www.sciencedirect.com/science/article/pii/S0925231212006509#
I would look at this formulation. You can use GP and other method of learning.
Conference Paper Plug-and-Play hyper-heuristics: an extended formulation
A good standard approach is based on Bayesian Decision Trees which are easy to handle an additionally give you the opportunity for an self-learning approach. You will find this topic in many standard textbooks on Machine Learning (e.g. cf. Alpadin: http://mitpress.mit.edu/books/introduction-machine-learning). It is at least a good starting-point.
Your context is not very clear to me. What's your objective?
For self learning intelligence in machine, you can refer machine learning concepts in ISLR and ESL books. Also, Andrew Ng's ML stuff may be useful.
I would look in using a form of Genetic Programming, with a hill-climbers. Perphaps you may consider to evolve some neural networks.
- Badges
- Science topic