ICIEOM2020_FULL_0007_37130
INFORMATION SECURITY AND MACHINE LEARNING: ENCRYPTION ALLOCATION BASED ON RECOGNIZING OF TEXT PATTERNS
ÁREA: 07. Knowledge and Information Management
RESUMO:
The productive processes have presented a high demand and generation of data and information. However, risks and attacks on the digital data of the companies have presented significant threats and losses to the industries that are embedded in the industrial revolution 4.0. Thus, this paper aims to present the feasibility study of an automated methodology, focused on cryptographic allocation, according to the degree of confidentiality of its content. Based on text pattern recognition algorithms such as Multilayer Perceptron (MLP) and Support Vector Machine (SVM), the proposed solution promotes a series of experiments in order to analyze whether certain information is confidential or ?very? confidential. After analyzing the information within the message content, using text pattern recognition techniques, the encryption algorithm responsible for encoding the message is chosen. With this approach it possible to choose either stronger encryption for very confidential information or a weaker encryption for not-so-confidential information. The experimentation has a performance analysis in order to evaluate the computational cost of the processes involved.
PALAVRAS-CHAVE:
machine learning, information security, cryptography, classified information
DOI: 10.14488/ijcieom2020_full_0007_37130
RESUMO:
The productive processes have presented a high demand and generation of data and information. However, risks and attacks on the digital data of the companies have presented significant threats and losses to the industries that are embedded in the industrial revolution 4.0. Thus, this paper aims to present the feasibility study of an automated methodology, focused on cryptographic allocation, according to the degree of confidentiality of its content. Based on text pattern recognition algorithms such as Multilayer Perceptron (MLP) and Support Vector Machine (SVM), the proposed solution promotes a series of experiments in order to analyze whether certain information is confidential or ?very? confidential. After analyzing the information within the message content, using text pattern recognition techniques, the encryption algorithm responsible for encoding the message is chosen. With this approach it possible to choose either stronger encryption for very confidential information or a weaker encryption for not-so-confidential information. The experimentation has a performance analysis in order to evaluate the computational cost of the processes involved.
PALAVRAS-CHAVE:
machine learning, information security, cryptography, classified information
DOI: 10.14488/ijcieom2020_full_0007_37130
AUTORES:
PEDRO HENRIQUE JUNIO SANTOS
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
CARLOS AUGUSTO DE FARIA COSTA JUNIOR
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
LUIZ MELK DE CARVALHO
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
FLÁVIO HENRIQUE BATISTA DE SOUZA
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
DIVA DE SOUZA E SILVA RODRIGUES
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
PEDRO HENRIQUE JUNIO SANTOS
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
CARLOS AUGUSTO DE FARIA COSTA JUNIOR
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
LUIZ MELK DE CARVALHO
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
FLÁVIO HENRIQUE BATISTA DE SOUZA
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
DIVA DE SOUZA E SILVA RODRIGUES
CENTRO UNIVERSITÁRIO DE BELO HORIZONTE UNIBH
|