Early detection of anomalous behavior of time series data is very important in many domains. The study deal with critical health event(CHE) that can lead to mortality in intensive care units of hospital. A novel approach which uses layered learning to carry out classification tasks for prediction of acute hypotension episode(AHE) or tachycardia episode(TE) is used. Layered learning denotes a hierarchical machine learning algorithm in which the entire prediction task is split into two or more layers. The learning process for a particular layer is affected by learning process at earlier layer. Pre-conditional events are the events which take place before the final events in which we are interested. In layered learning, initially modelling is done for the pre-conditional events. Based on pre-conditional events prediction for final events is obtained. Time sequences are distributed in three windows named observed, warning and target windows based for fixed time duration. According to health attributes in target window the event of AHE or TE is decided.
Input variables : Heart rate, systolic blood pressure, diastolic blood pressure, mean arterial blood pressure
Output Variables : Status of Acute Hypotensive Episode or Tachycardia Episode
| Statistical | : | Mallow's CP | R Squared | Mean Square Error | Adjusted R Squared | Mean Absolute Error | Huber Loss |
| Infrastructure | : | Log Bytes | Logging/User/IAMPolicy | Logging/User/VPN | CPU Utilization | Memory Usage | Error Count | Prediction Count | Prediction Latencies | Private Endpoint Prediction Latencies | Private Endpoint Response Count |
Visit Model : github.com
Additional links : arxiv.org
| Model Category | : | Public |
| Date Published | : | October, 2019 |
| Healthcare Domain | : | Provider |
| Code | : | github.com |
| Utilization Management |
| ED Admission |