Is it possible to make dynamic program instead of the static one ?
See, any program is static i can say because it works for something fix we coded for and ya that's why we can say it as 'static program'.
Any program is able to give fix output based on the given parameter values only.
Than in terms of machine learning, how can we create such programs that can create dynamic outputs based on dynamic inputs ? because for Machine learning algorithms, there may be always different types of inputs and we need to give output based on it.
That's the big question coming to my mind.
So again i am asking that, is it possible to make dynamic program instead of the static one ?
I'd say most of the programs are dynamic in nowadays.
i.e.: you just wrote post which contains some text, let's say A, I wrote mine post, which contain's text B. Surely nobody had to newly program and compile JForum software so we could post our posts with different texts.
So probably explain more clearly and maybe with some interactive examples what do you mean?
What's regarding machine learning, I still find this topic very fuzzy, even though I think it is floating around for the past 10 years, might even more.
Personally myself it is still difficult to distinct what is considered machine learning and what is not from practical point of view.
If to talk about some knowledge base which program populates and every time uses to define its next decision how to act - in this context I don't find anything fascinated or anything what wasn't around for some time.
Are you talking about that?
when in doubt - just merge, and never rebase
posted 9 months ago
See if we wrote one program and we are passing different values to it and we are considering that program works dynamically than I can say that it's wrong approach because we are using same content with different values only.
By the word dynamic program means, it can generate different output with different types of inputs given to it. Which can be totally new and also not of the same type which were previously given to it.
Because in machine learning, each time we will get different types of input and each time we need different types of output. There is no static algorithm to work behind the output ideally.
And that's why I want to know that is it possible to create such program which can operated differently as per the inputs given?
You can do it the object‑oriented way, but not in the beginning forum, so I shall move you.
Go through the Java® Language Specification (=JLS) and see if they haven't still got the example about arithmetic. Yes, here it is. In that case, you can pass a different instance to your code and get different results. You can consider expanding that concept and use machine learning to fill in methods in sutypes, and pass instances of those subtypes to your code. Have a look at the video Liutauras posted; that will probably help. You may need to recompile your new classes on the fly; I don't know how to do that.
At Coursera they are doing a course on machine learning. It looks very interesting. and I would have taken the first course, were it not that they are using Python and Numpy, so I first have to rehearse some Python. The price is $39 for a month, that compares to a decent book about the subject. Maybe worth a try.
Ruth Stout was famous for gardening naked. Just like this tiny ad:
RavenDB is an Open Source NoSQL Database that’s fully transactional (ACID) across your database