Predictive Classification with Imbalanced Enterprise Data

Daskalaki, Sophia; Kopanas, Ioannis; Avouris, Nikolaos

doi:10.1142/9789812779861_0003

Cited by 1 publication

(1 citation statement)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ird, at the same time, with regard to the characteristics of large data business types and large data volumes of large enterprises [13,14], business personnel cannot quickly identify which sensitive data is based on standards, and the level of confidentiality corresponding to sensitive data is not easy to define. erefore, establishing a classification and gradation protection catalogue for the corresponding trade secrets of large enterprises, establishing a more detailed classification and gradation process specification and implementation guide according to the business data, improving the efficiency and accuracy of classification and gradation management, providing reference for confidentiality of documents for business personnel, and forming compliance management measures are imperative [15,16].…”

Section: Introductionmentioning

confidence: 99%

Application Research of Intelligent Classification Technology in Enterprise Data Classification and Gradation System

Wang

Chang

et al. 2020

Complexity

View full text Add to dashboard Cite

Classification and gradation system adopts different security protection schemes for different types of data by implementing classification and gradation management of data, which is an important pretechnical means for data security protection and prevention of data leakage. This paper introduces artificial intelligence classification, machine learning, and other means to learn and train enterprise documents according to the characteristics of enterprise sensitive data. The generated training model can intelligently identify and classify file streams, improving work efficiency and accuracy of classification and gradation. At the same time, the differences, advantages, and disadvantages of K-NN (K-Nearest Neighbors), DT (Decision Tree), and LinearSVC algorithms are compared. The experimental data shows that LinearSVC algorithm is applicable to high-dimensional data, with discrete, sparse data features and large number of features, which is more suitable for classification of sensitive data of enterprises.

show abstract