matrix-1013611_1280

CM is collaborating with an international team of data scientists to develop and implement the Advanced Data Systems Analysis (ADSA), an algorithm to rank and cluster multi-dimensional (multi-sectoral) items. ADSA provides a wide range of advantages compared to other tools currently prevalent on the market and can be implemented in a variety of forms including consultancy and research papers, cloud-based software applications or as an algorithm that can be integrated into a company’s own data analysis section or automated industrial systems.

The areas of application include such fields as financial risk analysis, pricing optimization, product positioning, credit scoring, political risk analysis, or emotion scoring. Moreover, automated systems such as manufacturing and production processes can be equipped with the ranking and cluster algorithm.

ADSA was originally founded by a highly skilled team of data scientists consisting of Remi Mollicone, Ing. and Giovanni Feverati, PhD, who develop the algorithm, and Johann Barbier, PhD, and Maxime Buemi, MA, who help to implement it. The quality of our product and its implementation is ensured by the skills and experience of the individual team members, but also by the cross-functional cooperation between IT specialists, mathematicians and business experts.

Basically, ADSA is a set of the following modules:

  • Ranking tool which ranks items/objects based on weights that are automatically extracted from the data without human intervention
  • Clustering and interaction tool to cluster items or variables based on their similarities and dissimilarities and to calculate and analyze the interaction between these clusters
  • Visualization module to illustrate the ranking, the clusters and their similarities of the clusters
  • Composite indicator module that constructs indicators for each variable, dimension etc.
  • Complexity, stability and resilience module which provides information on the complexity of the data and to what extent the variables and items contribute to its complexity (currently under development)
  • Probabilistic network module which includes a graphical tool to represent the conditional dependencies between items and between variables (currently under development)


The advantages of ADSA

Our clients benefit from a wide range of advantages provided by ADSA modules, which go far beyond the conventional ranking tools.

The main advantages of ADSA modules are:

  • The holistic approach of ADSA, which can be applied in different industries and sectors dealing with a wide range of problems
  • The ability of ADSA to analyze highly complex and multi-dimensional datasets
  • The ADSA algorithm is able to produce an outstanding accuracy which can be quantified by indicators
  • ADSA rankings and clusters are based on weights which are automatically extracted from the raw dataset without human intervention
  • ADSA measures the contribution of each variable to the ranking
  • Modules have the ability to estimate the stability of the ranking
  • ADSA is able to calculate and illustrate the hierarchical structure of items and clusters by dendrograms and geodesic distances
  • ADSA is a cross-sectional tool and can extract more information from the data than many others, which helps to improve the learning process and to get better results
  • ADSA is also able to deal with pooled data which combines cross-sectional and panel data
  • Customized implementation according to the client’s needs


Industries and sectors

With the growing amount of data the expectations of actors in different industries and sectors to analyze and use these data have gone sky high in recent years. ADSA and its different modules were designed to analyze complex data and to model sophisticated problems, which makes ADSA a highly valuable product that can solve a wide range of problems.

On the one hand, ADSA can be used by companies, governments or organizations for the purpose of information gathering. The modules allow our clients to rank and cluster items and objects, which can help them in making difficult decisions based on objective information. This includes fields such as financial risk analysis, pricing optimization, product positioning, credit scoring, political risk analysis, or emotion scoring. On the other hand, the ranking and cluster algorithm can be implemented into automated manufacturing and production processes and other automated systems, a field in which rankings are very common.


The implementation of ADSA

CM works with its partners to implement ADSA in a variety of forms, including:

  • Consultancy (short-term)
  • Research papers (short-term)
  • Cloud-based software tools (long-term)
  • Integration of algorithm into
    • analysis section and IT architecture of a company (long-term)
    • automated production and manufacturing processes (long-term)

If our clients wish to implement ADSA as long-term modules into their own analysis section or into a manufacturing process, our team can configure ADSA for automated use according to the client’s needs.

The process of implementation involves four service-intensive stages:

  • The implementation process starts with consultancy and the harmonization of the client’s problem and our solution.
  • In the second phase of the implementation our team configures and adapts our solution to the client’s problem.
  • The third stage includes the actual implementation of the algorithm such as the installation, consultancy and teaching
  • As soon as the implementation of ADSA is completed our team is able to provide further support and updates


Case study

If you are interested in how ADSA works, have a look at the ADSA case study on natural disasters.

 

One comment

Comments are closed.