Integration of Fuzzy Inference Systems and Random Forest for Intelligent IoT-Based Health Monitoring

Abstract –Healthcare has become a major global issue due to
the recent corona virus outbreak, highlighting the importance
of effective and scalable solutions such as health monitoring
systems based on the internet of things. Smart phones and
wearable sensors have greatly contributed to the rapid
expansion of the IoT. The ability to remotely monitor health has
greatly improved diagnosis accuracy and halted the spread of
infectious diseases. This paper introduces a health monitoring
architecture driven by the IoT. The architecture employs data
pre-processing to improve analysis by removing duplicates,
changing inputs, and standardizing records. For feature
selection, the employ the ERFE method, and for feature
classification, they turn to an ensemble-based machine learning
approach. The proposed model, FIS, employs fuzzy logic, fuzzy
sets, and bootstrapped sampling to enhance the IFDT random
forest ensemble, making it more robust and long-lasting across
various healthcare data sources. The FIS-DT model
outperforms other approaches, according to experimental
evaluation, which reaches an accuracy of 94.43%. These
findings demonstrate that smart monitoring systems built on the
Internet of Things have the potential to enhance public health
readiness before, during, and after a pandemic by providing
trustworthy remote healthcare solutions.
Keywords—Clinical Decision Support System (CDSS),
Enhanced Recursive Feature Elimination (ERFE), Structural
Health Monitoring (SHM), NDE (Non-Destructive Evaluation