NxNeural

New Product

Data Mining for Marketing
30 years of Experience in Applied Mathematics
(neural networks, data analysis, etc.)
30 years of Projects: Defense, Road Safety, Automotive Marketing.
A team that teach at Engineers Schools

 

Nx Neural Product

Nexyad offers NxNeural, a new AI Tool integrated into Excel in the form of a menu, ready for use. Nx Neural uses all the power of Excel for data management and its graphics for display. Data Scientists can easily analyse Data on Excel without coding with Python or C++, for Data Manipulation function and neural processing function:

. Factor Analysis
. Statistics
. Deep Learning
. Fuzzy Logic
. Neural Gaz

We give our customers answers to their questions:

  • Customer Segmentation
  • Forecast of Activity and Turnover
  • Analysis of Satisfaction Surveys
  • Scoring
  • Detection and Explanation of Attrition
  • Specific Questions: linking Data to explain or predict a Phenomenon

 

 

Expertise and Know-How

Pre-Analysis

We bring full support in preparing Data for analysis.

  • Coding of computer fields into mathematical variables
  • Processing unfilled fields
  • Adjustment of variables to move from sample statistics to market statistics

 

Analysis Methods

We use the best methods for Data analysis

  • Neural regression: Non-linear modeling of a function
  • Neural discriminant analysis: non-linear “A.D”
  • Restricted component analysis: neuronal “PCA”
  • Neural guidance: reduction of redundancy under constraints
  • Classification: neurons gas
  • 2D topological mapping (VQP)

 

 

Example of Application

STELLANTIS (former PSA Group)

Automotive Customer Segmentation
for City Cars Low Range

  • Inputs are a mixture of quantitative and qualitative data: automobile trends, prices, technical data, socio-cultural data, model previously purchased
  • Output: classes of homogeneous buyers best segmenting the market with complete typology of buyers.

This segmentation was used for the specification of PSA products.

VQP mapping example for City Cars low segment