EVIS delivers services to transform your data to information.

Data Mining is often considered to have its roots and to be a blend of many branches of science.

Why would you choose for us?

EVIS gives advice and provides services that lead to improvement of a corporation's marketing, sales, and customer support operations through a better understanding of their customers and their needs. EVIS uses its fundamental and deep expertise in order to process, analyze and interpret data from different fields: banking, telecommunication, e-Business, retail industry, Medical Care, Public Sector, Oil exploration, Engineering and Semiconductor industry.

Our expertise is based on a thorough understanding of many Data Mining methods and tools. In contrast to our competitors, we do not only use well-known commercial Data Mining tools like SPSS and SAS. The reason is that these tools are typically used as "black boxes" for data exploration or knowledge discovery, but give you less freedom and power to explore the data in depth. Using our well-defined Data Mining strategy, we transform your data into valuable information by developing tailored on-the-case Data Mining solutions. Based on our expertise and experience, we combine several techniques that are best for a particular situation. Clearly, this depends on the nature of the Data Mining task, on the available data and on the business goal!

EVIS has a broad expertise in project management, systems and requirements engineering on the one hand, and combines this successfully with our deep technical understanding in mathematics and computer science. This allows us to succesfully carry out projects that solve different business problems.

We invite you to take a look at the advantages of applying Data Mining solutions in different branches.

In what follows we show why our expertise in understanding your business goal, mathematical statistics, algorithm development and application are of vital importance for successfull projects.

For your company Data Mining means discovering gold nuggets, but only if you know how to apply it!!

Why do we say this? Because Data Mining practitioners are prepared to use ad-hoc methods only if it can be shown that they work, even if they do not meet the mathematical rigor. In Data Mining, more experimental work is done: looking at the same problems, Data Mining practitioners have found solutions that work well, even if they cannot be proven, as rigorous statisticians would require.
We illustrate this by means of one example: the Statistical Inference is not considered in Data Mining applications. This can lead to wrong conclusions and decisions - as in the example below.

Example for blind application of Data Mining: there is a famous illustration of the statistical inference importance known as “Black Swans: expecting the unexpected” [Nassim Taleb]: “No amount of observations of white swans can allow the inference that all swans are white, but the observation of a single black swan is sufficient to refute that conclusion.” Blind application of the algorithms might be mining data to confirm whatever 'white swans' that one hopes will provide business profits. But what if a black swan happens?

Some Concluding Remarks

 Data Mining is looking to a business problem fundamentally in a different way than Statistics:

• The classic statistical methods require that a model/hypothesis is given beforehand and then that the data is used to test the model in order to accept or reject it. In statistics the model is playing the central role. The computation and the model selection criteria are secondary and used for other reasons than in Data Mining.
  Data Mining algorithms use the data to come up with useful models for a user. The number of models that can be induced from a set of data is huge; the current Data Mining algorithms are able to handle this due to the computational power of the computers today. The algorithms are more central in Data Mining because of the computer science and other related disciplines that are important branches within Data Mining.
• Data Mining algorithms often use data from different databases and this complicates the models induction. These databases might have noisy and contradictory data and they might be setup for other reasons than data mining.
  In statistics, data is often collected for the purpose of testing hypothesis: the focus is on gathering of relevant and useful data.
• The statistical packages are very efficient in handling numerical data but poor in handling other types of data like the categorical ones.

It is very important that:

• Data Mining practitioners have to be aware of rigorous statistics methods and the behavior of the Data Mining algorithms when applied to specific data. 
• Data Mining practitioners have to understand the business goal, what kind of data and which algorithms they have to consider in order to reach their scope.
• Many people use the term Intelligent Data Analysis (IDA) and Data Mining synonymously since most of the IDA techniques are applied in Data Mining.