Enhanced Over_Sampling Techniques for Imbalanced Big Data Set Classification

Improved classification of large imbalanced data sets using rationalized technique: Updated Class Purity Maximization Over_Sampling Technique (UCPMOT)

“…1 demonstrates the organization of work structure to apprise imbalanced Big Data sets [36]. The work involves the investigational analysis of over sampled data set to improve classification.…”

“…The nonbinary-class data sets are pre-processed using either traditional method (One versus One (OVO) and One versus All (OVA)) or the suggested LVH method. LVH [36] overcomes the disadvantages of traditional methods improving classification results. This method considers only the highest majority class versus each of the minority class satisfying I.R.…”

“…The stated technique (UCPMOT) works with either of the four basic techniques specified below, for elementary over_sampling process [36]. They are classified into two categories viz.…”

Nuevos Algoritmos Basados en Grafos y Clustering para el Tratamiento de Complejidades de los Datos

“…Subhash et. al [82] presentan técnicas de sobre-muestreo, todas estas desarrolladas con técnicas de clustering. La primera propuesta se denonima CME (Clustering Minority Examples), la cual sólo involucra las instancias de clases minoritarias para generar instancias sintéticas a través del algoritmo K-Means.…”