Growth optimization-based stacked bidirectional long short-term recurrent neural network model for detecting breast cancer in Internet of Things healthcare environment

The healthcare industry is the most prominent user of Internet of Things (IoT) technology. In the healthcare monitoring framework, IoT devices give data about individual patients. In addition, individuals can check their health using smart devices, so IoT is a crucial part of the healthcare administration system in general. Breast cancer (BC), which is caused by the abnormal and rapid development of breast cells, is the most prevalent cancer in women. Early recognition of malignant cells is essential for reducing cancer-related mortality. Patients with BC can benefit greatly from prompt diagnosis and treatment. To overcome the challenge of detecting BC in its early stages, this study presents a medical IoT-based diagnostic scheme based on a stacked bidirectional long short-term recurrent neural network (SBLRNN). Machine learning algorithms rely heavily on hyperparameters, since they have direct control over the behaviors of training algorithms and have a major impact on how well those models perform. To improve SBLRNN's classification performance, this research uses a growth optimization algorithm to choose datasets with better features. The proposed process was put to the test by means of the BC Wisconsin (Diagnostic) dataset. Classification accuracy for the suggested model was 96%, while the area under the curve score was 98%.