Research on Embedded Multifunctional Data Mining Technology Based on Granular Computing

Area of Science:

• Computer Science
• Data Science
• Artificial Intelligence

Background:

• Multisourced, heterogeneous, and unbalanced embedded multifunctional data present challenges for current data mining technologies, leading to low accuracy.
• Existing data mining methods struggle with the complex characteristics of embedded multifunctional data, limiting their effectiveness.

Purpose of the Study:

• To develop an improved embedded multifunctional data mining technology using granular computing.
• To address the limitations of existing data mining techniques in handling complex data characteristics.

Main Methods:

• Preprocessing embedded multifunctional data through reduction, standardization, and balancing.
• Applying data granulation to preprocessed data, calculating characteristics like offset, particle density, and intraparticle interval.
• Utilizing a neural network with granular features for data classification, anomaly detection, and fault identification.

Main Results:

• The granular computing approach demonstrated high accuracy in data mining tasks.
• Anomaly mining results across different data types exceeded 0.9, confirming the technology's effectiveness.
• The proposed method successfully performed data classification, anomaly detection, and fault identification.

Conclusions:

• Granular computing offers a robust solution for embedded multifunctional data mining.
• The developed technology significantly improves accuracy and effectiveness in data analysis.
• This approach overcomes the limitations of traditional data mining for complex, heterogeneous datasets.