Bi-LSTM based model to predict open ion-channels
Many diseases, including cancer, are believed to have a contributing factor in common. Ion channels are pore-forming proteins present in animals and plants. They encode learning and memory, help fight infections, enable pain signals, and stimulate muscle contraction. If scientists could better study ion channels, which may be possible with the aid of machine learning, it could have a far-reaching impact.
When ion channels open, they pass electric currents. Existing methods of detecting these state changes are slow and laborious. Humans must supervise the analysis, which imparts considerable bias, in addition to being tedious. These difficulties limit the volume of ion channel current analysis that can be used in research. Scientists hope that technology could enable rapid automatic detection of ion channel current events in raw data.
In technical terms, this is a classification problem. The features (X) is the signal values and the target vector (y) is open_channels. The distribution of target vector open_channels is imbalanced.
raw_data is obtained from here : https://www.kaggle.com/c/liverpool-ion-switching/data
data-without-drift is obtained from here : https://www.kaggle.com/cdeotte/one-feature-model-0-930/data
This repository is based on kaggle competition called 'liverpool-ion-switching' (https://www.kaggle.com/c/liverpool-ion-switching).
Jupyter notebook Remove_Drift_Tomek_Links_BILSTM.ipynb contains the finalized model. The final model used Drift-removal + Undersampling + Bidirectional LSTM approch, which resulted in the 93.95 F1-Macro score. The performance on individual classes can be seen in the confusion matrix.
Model specifications are given below (after hyperparameter tuning) :
| Method | Numbers |
|---|---|
| LSTM | 3 layers |
| Droput Value | 0.5 |
| Bidirectional | 3 layers |
| Optimizer | adam |
| Class reweighing | yes (μ=0.15) |
| Undersampling | yes (TomekLinks) |
| Sequence Length | 1000 |
| Nodes | 256 |
| Batch size | 128 |
| Learning Rate | 0.001 |
| Accuracy Score | 0.969 |
| Precision Score | 0.939 |
| Recall Score | 0.939 |
| Final Submission Score (F1-macro) | 0.939 |