Implementation LSTM Algorithm for Cervical Cancer using Colposcopy Data

Cervical cancer ranks second highest cause of death in women in various worlds. This happens because most women are not aware of the symptoms of cervical cancer in the early stages. To reduce the number of deaths caused by cervical cancer by identifying the symptoms of cervical cancer in the early stages. Identification of early symptoms of cervical cancer can be made with colposcopy tests that produce colposcopy image data. Colposcopy test is a method to identify cervical cancer based on images of the cervix with an enlargement of up to 10 times and it gets accurate results. Accuracy results from colposcopy tests can be improved by using computational calculations. Besides being used to improve accuracy, computational calculations also make it easier for people to detect cervical cancer. In this study, computational calculations are performed by implementing the Long Short-Term Memory (LSTM) algorithm to identify cervical cancer using colposcopy data. The implementation of the LSTM algorithm in the classification process of colposcopy data with an optimal number of hidden layers of 150 hidden layers results in an accuracy rate of 66%.