Identification of marker features in multi-dimensional data

A feature identification system for identifying marker features of one or more subsets of a multi-dimensional data, each subset including a plurality of data points associated with a plurality of features, wherein each of the plurality of data points defined by feature values corresponding to the associated features, the feature identification system comprising:

• a processor;
• an interface;
• and
• a memory coupled to the processor, the memory comprising:
• a computation module configured to,
• identify a plurality of feature pairs based on a matrix between the plurality of data points and the plurality of features;
• compute correlation distances between features in each of the plurality of feature pairs based on correlation co-efficient values, wherein a lower correlation distance between features indicate a highly related feature pair;
• and
• generate a semi-circle pattern of the plurality of features from a linear pattern of the plurality of features, in a two-dimensional plane, based on the correlation distances, such that the features in the highly related feature pair are placed close to each other than the features not highly related to each other;
• evaluate force exerted on each of the plurality of data points by the each of the plurality of features and position of each of the plurality of data points within a region defined by the semi-circle, in the two-dimensional plane, based on the evaluation;
• and
• a rendering module configured to,
• calculate a mean feature value for each of the associated features of the data points of a particular subset from amongst the one or more subsets, wherein based on the calculation, an overall mean of the mean values associated with each of the plurality of features across each of the plurality of data points is identified;
• represent the overall mean feature value as another semi-circle in the two-dimensional plane and with respect to the overall mean feature value, represent the mean values of each of the features corresponding to the data points in the particular subset in the two-dimensional plane, wherein the representing comprises identifying a lower bound and an upper bound corresponding to each feature for the particular subset for being plotted in the another semi-circle in the two dimensional plane;
• and
• determine the marker features based on a relative position of the mean feature values with respect to the overall mean feature value in the two-dimensional plane, wherein the marker features distinguish a first subset of the multi-dimensional data from a second subset thereof.

A non-transitory computer-readable medium having embodied thereon a computer program for executing a method for identifying marker features of one or more subsets of a multi-dimensional data, each subset including a plurality of data points associated with a plurality of features, wherein each of the plurality of data points is defined by feature values corresponding to the associated features, the method comprising:

• identifying a plurality of feature pairs based on a matrix of the plurality of data points and the plurality of features;
• computing correlation distances between features in each of the plurality of feature pairs based on correlation co-efficient values, wherein a lower correlation distance between the features indicate a highly related feature pair;
• generating a semi-circle of the plurality of features from a linear pattern of the plurality of features, in a two-dimensional plane, based on the correlation distances, such that the features in the highly related feature pair are placed closer to each other than the features not highly related to each other;
• evaluating, by the processor, force exerted on each of the plurality of data points by each of the plurality of features and positioning each of the plurality of data points within a region defined by the semi-circle in the two-dimensional plane, based on the evaluation;
• calculating a mean feature value for each of the associated features of the data points of a particular subset from amongst the one or more subsets, wherein based on the calculating, identifying an overall mean of the mean values associated with each of the plurality of features across each of the plurality of data points;
• representing the overall mean feature value as another semi-circle in the two-dimensional plane and with respect to the overall mean feature value, representing the mean values of each of the features corresponding to the data points in the particular subset in the two-dimensional plane, wherein the representing comprises identifying a lower bound and an upper bound corresponding to each feature for the particular sub-set for being plotted in the another semi-circle in the two dimensional plane;
• and
• determining the marker features based on a relative position of the mean features values with respect to the overall mean feature value in the two-dimensional plane, wherein the marker features distinguish a first subset of the multi-dimensional data from a second subset thereof.