I came across few articles that emphasis on big data analysis with java, scala and python and High performance computing with c, c++, FORTRAN (and their flavors) but I want to ask is their any language that supports both? and what are the features required for any language to support both?
There are many ways to model big data such as time series, Bayesian, simulation, etc. The use of software depends on both the research objective and one's ability to understand the programming language.
Hi,
You can used paython, Java Enabled API with Weka, R and RapidMiner. Its depends on you. You are welcome and i will guide you how to fix it with specific tool.
Review my blog for Apache Hadoop and Java!
http://hadoopkickstart.blogspot.de/2014/02/hadoop-kick-start-hadoop-single-click.html
Regards
Muddsair
The advantage of Python are the plethora of available libraries for data analysis and machine learning.
For High-performance computing there's hardly a way around using a language that is faster, like C, C++ or Fortran.
The practical way to go is to use Python for the overall approach, and use C or Fortran for the computationally intense bits. This is what all the python machine learning libraries do, in fact.
Having said that, there are new dynamic languages coming up that rival python in useability while at the same time competing with C and Fortran in terms of speed. Look at Julia, for example. However, Julia still lacks far behind python in terms of choice of libraries.
So practically, the way most people in data science go is probably using Python for comfort and wealth of features with C/Fortran underneath for speed.
Why use one language?
One way is to assess the question whether it should be functional, procedural, object-oriented or aspect-oriented. (align with your modeling method).
Another way is performance, then assembler is best - you will get no unnecessary computations - but it will be quie tricky to read the code.
Yet, if there is a nice toolkit IDE just use it.
What is your requirements? one language to rule them all, or performance, or good components, or a large user base or ...
For big data before the programming, you have to decide the storage and the processing parts separately. This will make you to decide the language on which you could better implements that.
You can also consider Java[Map/Reduce] on Hadoop platform for big data processing.
This is more a matter of the programmer and the way the data are handled than the language itself...!
Mostly, I agree with the prior comments. When deciding on the language, it is a matter of preference. It depends on how much resource there is and how the data is going to be processed. Some mining algorithms are more computationally expensive than others. Or, there may be a lot of records but, only a few attributes are relevant.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python