Improved Efficiency of NoSql using Optimized K-Mode

Abstract

Data Mining is the process to extract unknown, valid patterns and relationships that provide useful
information. Non relational databases are a broad class of database management systems identified by non-adherence to
the widely used relational database management system model. Non relational databases are not built fundamentally on
tables, and generally do not use SQL for data manipulation. Non relational database systems are often highly optimized
for retrieval and appending operations and often offer little functionality beyond record storage. The reduced run-time
flexibility compared to full SQL systems is compensated by marked gains in scalability and performance for certain data
models Clustering technique defines classes and put objects which are related to them in one class on the other hand in
classification objects are placed in predefined classes. There are various algorithms used for clustering. These are k
means algorithm, k-medoid algorithm, k-nearest neighbour algorithm, k-mode algorithm etc. K-Modes algorithm is an
extension of K-Means and uses simple matching dissimilarity function instead of Euclidean distance. The major
drawback of k-mode is that the user needs to define the centroid points. The nature-inspired harmonic algorithm is
hybridized to optimize the k-mode algorithm. Harmonic K-Mode Algorithm is proposed in this work that reduces the
computation time and improves the execution time for cluster generation.