In biomedical applications, time series data is frequently observed along structured information. ShortFuse, an approach where accuracy of deep learning models boosts for time series by considering structured covariates such as patient demographics and measures from clinical examinations in the dataset. ShortFuse outperform in two biomedical applications regarding osteoarthritis and cerebral palsy. It uses hybrid convolutional that incorporate the covariates via weights that are shared across the temporal domain. ShortFuse introduces specialized structure called 'hybrid layers' for fusing structured covariates with time series data. The hybrid layers use structured information as distinct inputs, which are used to parametrize, guide and enrich the feature representation. In case of osteoarthritis, the aim was to predict the progression of osteoarthritis, in terms of an objective measure of cartilage degeneration called Joint Space Narrowing. For cerebral palsy, the task was to predict whether the psoas lengthening surgery will have a positive outcome. The surgery is procedure to address a tight or overactive muscle in the pelvic region.
Input variables : Patient Demographics, Time Series data from accelerometer
Output Variables : Osteoarthritis Prediction, Prediction of postive outcome for cerebral palsy
| 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 | : | September, 2019 |
| Healthcare Domain | : | Provider |
| Code | : | github.com |
| Health Risk Management |
| Chronic Care Management |