Recasting Java neural networks in Python

Many neural network applications implemented in Java, such as Neuroph, Encog and Joone, may look rather different when switching from the Java language  to Python with the help of  the DMelt computing environment. First of all, they look simpler. You can use your favorite Python tricks to load and display data. The Python coding is simpler for viewing and fast modifications. It does not require recompiling after each change. At the same time,

Coding graphs for data mining in Python using Java platform

Graphs belong to the field of mathematics, graph theory. For data analysis that requires searches of particular patterns, graph-based data mining becomes an important technique. Indeed, in real life, most of the data we have to deal with can be represented as graphs. A typical graph consists of vertices (nodes, cells), and of edges that are the connecting lines between the nodes. Analysis of graphs includes determining certain details about the graph structure. For example, determining all routes or the shortest paths between two nodes or cells.

Data scientists, spy-statisticians

Have you ever thought why the data scientist profession is so hot right now? You are proboably familiar with the answer. Data science is a hot commodity because of large data volumes and new computerized techniques. However,  have you thought why this data value is so large now, compared to, say 2000? And this is a good question to answer. We will attempt to do it in this blog.

Popularity of programming languages in 2016

According to TIOBE Index for January 2016, the Java popularity index again has reached 21%, leaving behind C++ (6%), while Python index is only 3.8%. These numbers can be different for data analysts positions, of course, where Python is likely to be more popular than Java.

JTerm - Linux commands for Windows?

JTerm is an attempt to mimic Linux/Unix commands using 100% Java. Thus it works on Windows. This project is its infancy, but it is already very useful. For some, it may also look also more attractive than using Cygwin on Windows. First thing to notice - it is only 3M in size, a tiny program compared to the default Cygwin installation. Secondly, it has nice look and feel thanks to Nimbus L&F. In combination with JPort portable Java desktop, it may lead to interesting alternative to Cygwin.

5 best free pograms for statistics, data mining and graphics

R-package - a software for statistical computing written in C. Script oriented. Pros: widely used, simple, extensive documentation. Cons: simpler graphics compared to competitors, no multi-threading, scripting is less powerful compare to other programming languages.

Java is slow? Java can read data faster than C++!

Java can be rather fast! According to the article posted in ArXiv (, Java is faster than C++ in reading data (to be exact, data in the ProMC format, which is based on Google's Protocol Buffers). See Table I of this article.

"Small data". Never heard this term?

We often hear the term   "big data" (see "Big data" wikipedia link). Taking the path of finding cool words for description of something quite trivial (before we had "not quite big data", and in 10 years from now we will have "monstrously big data"?), how about a new term "small data"?  The definition of "small data" is less ambiguous than for "big data": Data that has small enough size for human comprehension.

jPort - desktop menu to access free portable Java applications

jPort is finally out (and it is already release 1.4!). What is special about jPort project? jPort is a portable application launcher for Java-enabled platforms (Windows, Linux, Mac). It can launch dozens of free Java-powered programs for office, science, education, code development, entertainment and graphics. If you need a tool to organize all your applications in one place, a single menu (for example, on menuless Windows8) - this is the way to go.

SCaVis. A successor of jHepWork

SCaVis is a successor of jHepWork. Why this change in its name? jHepWork has its origin in high-energy physics (remember, Higgs?). The "Hep" part of jHepWork abbreviates "High-Energy Physics". jHepWork is used in several HEP areas, but it is not very popular in this field since high-energy physics is almost completely based on the ROOT C++ package developed at CERN. The reason is that experimental data are written and stored in the ROOT format.


Subscribe to RSS - blogs

Guest Blog

Guest blog