|  | 0 days earlya | 3 days earlya | 6 days early | 9 days early | 12 days early |
|---|
Coxb | 0.955 ± 0.06 | 0.992 ± 0.02 | 0.870 ±0.01 | 0.85 ±0.01 | 0.810 ±0.01 |
k-NNc | 0.950 ± 0.02 | 0.909 ± 0.01 | 0.890 ±0.02 | 0.840 ±0.02 | 0.816 ±0.01 |
SVMd | 0.969 ± 0.04 | 0.954 ± 0.02 | 0.930 ± 0.03 | 0.895 ±0.04 | 0.857 ±0.02 |
DTe | 0.974 ±0.01 | 0.959 ± 0.03 | 0.924 ±0.00 | 0.897 ±0.01 | 0.869 ±0.03 |
BPNNf | 0.980 ±0.02 | 0.954 ±0.05 | 0.933 ±0.0 1 | 0.894 ±0.02 | 0.878 ±0.03 |
PNNg | 0.985 ±0.01 | 0.961 ±0.02 | 0.940 ±0.0 2 | 0.889 ±0.02 | 0.889 ±0.02 |
RNNh | 0.985 ±0.01 | 0.960 ±0.01 | 0.931 ±0.00 | 0.910 ±0.02 | 0.871 ±0.01 |
LSTMi | 0.990 ±0.01 | 0.961 ±0.02 | 0.937 ±0.02 | 0.920 ±0.03 | 0.897 ±0.03 |
T-LSTMj | 0.997 ± 0.00 | 0.969 ± 0.01 | 0.947 ±0.03 | 0.921 ±0.03 | 0.914 ±0.02 |
- a n days early: The models make prediction n days before the final death/survival time. They use sequence data from day 0 to n days before the last time to predict
- b Cox: Cox’s proportional hazards regression model is semi parametric regression model. It can analyze the influence of many factors on outcomes. It is used in [19]
- c k-NN: k-Nearest Neighbors method makes prediction based on the information of nearest k samples in training set. In this mortality prediction task, the most accurate results appeared when k = 3
- d SVM: Support Vector Machines classify by solving the separation hyperplane which can divide the training data correctly and has the largest geometric interval
- e DT: Decision tree is a simple classifier consisting of sequences of hierarchically organized binary decisions. It is used in [33]
- f BPNN: Back Propagation Neuron Network makes the signal and the error propagate forward and backward separately. It is used in [20]
- g PNN: Probabilistic Neural Network is a forward propagation network and does not need back propagation to optimize parameters by using Bayesian decision-making. It is used in [21]
- h RNN: Recurrent Neural Network have been introduced in the ‘T-LSTM’ section
- i LSTM: Long Short-Term Memory which we have introduced in the ‘T-LSTM’ section. Here, the hyperparameter setting is same as T-LSTM
- j T-LSTM: Time-aware LSTM is the model used in this paper. Its inputs are the three-dimensional vectors and the time intervals. The values for each dimension are the values of LDH, lymphocyte and hs-CRP in patients’ blood tests. Its output is the binary result 0/1. Here, 0 indicates survival and 1 indicates death. The hidden states in its units are 64 dimensional, and the fully connected layer has 32 dimensions
