Data Mining

 

 

 

 

Home    About Snape Signals      Signal Processing      Pattern Recognition      Data Mining      Research Papers

 

 

 

·        Data Visualisation and Analysis

 

One of the key tasks in mining data is the discovery of hidden relationships, similarities, and differences between records in the data set. Plotting the data records as points in a two-dimensional map often renders these relationships immediately obvious to a human. Difficulties arise because data is usually of high dimension and because the appropriate distance metric to be used in the map is not known: it is the choice of the right distance metric, which determines whether important hidden relationships are made obvious in the data map. Snape Signals has developed tools for projecting high dimensional data into two-dimensional maps and allowing different distance metrics to be rapidly applied. The tools have been used for applications ranging from analysis of customer creditworthiness, to analysis of cell phone base station traffic. For more detail see: Data Visualisation Tools (228kB) , Pre-processing and Data Mining (177kB) , and  Data Mining Papers (21kB)

 

 

·        Cellular Phone Traffic Prediction

 

The growth in cell phone use has been dependent on rapid network roll-out and expansion over the World. New base stations need to be positioned and commissioned at optimal positions and times to provide good service against capital investment. Snape Signals has used time series analysis techniques and data mapping methods to produce customised traffic prediction tools for cell phone networks.

 

 

 

 

Home      About Snape Signals      Signal Processing      Pattern Recognition      Data Mining      Research Papers