Datamining and clustering tool pros

What are the main perSimplex benefits?

  • Simple and user-friendly interface
  • Excellent processing speed
  • Ability to visualize millions of specific curves of numeric data
  • Input data may also contain mutually heterogeneous data
  • Ability to identify suitable clusters
  • Ability to distinguish level of similarity
  • Option to manage the quality of outcomes
  • Ability to identify abnormalities of input data
  • Non-hierarchical cluster algorithm
  • Linear computing complexity

A simple and user-friendly interface

perSimplex solution consists of two applications, SimplexDivide and SimplexImpera. Both applications are very simple to operate and are characterized by transparent functions. SimplexDivide application processes raw data according to parameters set up by business user and, consequently, they are visualized by SimplexImpera.

An excellent data processing speed

The speed of data processing is one of the greatest perSimplex software advantages and is related to the linear computing complexity of perSimplex algorithm. Moreover, SimplexDivide operates with data in second-intervals, thus generating space for immediate analysis of business challenges and flexible management in favour of your business.

The ability to visualize thousands of individual curves of numeric data

perSimplex does not lag behind in data visualisation. The effective and quick data processing by SimplexDivide is supported by presentation application SimplexImpera, which allows a business user to solve another task with identified clusters – valuable pieces of information.

The ability to analyze heterogeneous data

The input data can express not only the dynamics of time change, but they may as well include non-homogeneous data. It is sufficient enough to properly present numerical data and to take advantage of perSimplex. The cluster generation criterion becomes multidimensional thus allowing to deal with tasks related to the factor analysis.

The ability to identify adequate clusters

perSimplex is able to identify specific clusters within a context of data. A business user does not have to specify any final amount of clusters, nor he has to describe or limit the outcomes. These tasks are solved by perSimplex and it always considers the context of analysed data, the aim being to follow their natural structure.

The ability to distinguish the level of similarity

Another substantial advantage of perSimplex product is the ability to distinguish even very small differences in curve shape. This advantage can be considered as crucial, because the natural criterion for creation to create clusters is just the shape similarity of curves. It is just the shape similarity for curves, which is natural criterion for creation of clusters. Mutually more or less identical shapes of curves represent almost identical information.

The option to manage quality of outcomes

Business user has an option to set the way in which the similarity level of curves is evaluated and, moreover, he sets up a parameter whose value results in higher or lower strictness of curves shape evaluation. Higher strictness level causes generating a greater amount of clusters in which the curves are mutually more similar. As a result of increasing strictness, in general, we gain more detailed analysis. In principle, it is impossible to set an optimal level of strictness or of the inner similarity of curves in clusters. It depends on the intention of user with respect to analysed data.

The ability to identify abnormalities of input data

Real raw data usually contain some data abnormalities or non-standard data. Beside the ability to identify adequate clusters, perSimplex is also able to identify both too diverse and non-standard data. perSimplex identifies these data by creation of individual clusters.

In general, the ability to identify data abnormalities is useful in various practical cases, for instance, when an analyst is interested more in abnormal and non-standard data clusters than in dominant clusters. For example, either during the controlling and purifying of input data, or in the field of inspection and fraud detection.

Compared to various statistical techniques, this advantage of perSimplex is extraordinary since an analyst does not have to specify how abnormal data should look like or why they are abnormal. In addition, when using some statistical techniques the abnormal data are being lost and are labeled as statistically insignificant.

The non-hierarchical cluster algorithm

Being a non-hierarchical cluster algorithm, perSimplex is able to identify adequate clusters at each level of strictness independently, without being determined or limited by separation at a lower or higher strictness level.

The linear computing complexity

perSimplex product is characterized by the algorithm of linear complexity meaning that processing time increases continual proportion with the amount of curves and their points, and with the amount of identified clusters. A practical consequence in an ordinary computer is that the processing time of huge data amount takes only a few tens of seconds.