CloseClinical provides not only a portal of news and articles about health informatics, but also a discussion forum. More over, it has a vast directory of useful links and addresses, jobs and books.

SPSS Clementine is one of the very first softwares I tested for data mining. In fact, I began data mining with Clementine during at a time where not more than a few solutions were available. One can consider SPSS Inc., the company that produced Clementine, a pioneer in this regard.

Although I don't like commercial software in general, nor I like to write about commercial software on this website, it is by far the most straight forward and easy to use software that I found for effecient data mining. Let's revise its advantages:

1. There is a beautiful and ergonomic graphical user interface which does not bug every now and then, like many other beautiful GUIs do.

2. The software is shipped by a predesinged method -CRISP-DM- that I find both simple and efficient. The software do not impose the CRISP-DM method to the user. Rather, it helps the beginners with a good method and it can be considered a good educational initiative from this point of view. 3. The software proposes connections to different kinds of data source including SPSS exported files, excel, text, databases, etc.

4. The GUI imploys data flows for guiding the use of different subdivisions of the sample data through the experience. The advantage of data flows is that one do not need to reconstruct all the steps necessary for the preparation of data. They enable the user to view all experimentations on the same screen for further comparison of models.

5. In terms of modeling, it proposes a wide arsenal of modeling methods which include neural networks, linear and logistic regression for prediction; rule induction, regression and decision trees for classification; Kohonen networks, K-means and Two Step clustering for segmentation; APRIORI, GRI, and GRAMA for association; and Capri and rule induction for time sequence modeling techniques.

6. While the number of proposed models is not as high as those proposed by RapidMiner, they seem to me more efficient because of the extensive use of commercial and enhanced models. For example the C5.0 algorithm in Clementine, is an enhancement of Quinlan's C4.5 algorithm for decision trees which accepts both numeric and polynominal variables unlike many other non-commerical decision tree models.

If you work with different kinds of data, and you need a higher degree of reactivity in your data mining projects, or if you want to teach data mining methods in a university, or if you simply do not enough computer knowledge to code in R or SAS (in fact, if you prefere to code, you can do it with Clementine also), Clementine would definitely be a good choice for you.