=Paper=
{{Paper
|id=Vol-2563/aics_40
|storemode=property
|title=Interpretable Machine Learning Models for Assisting Clinicians in the Analysis of
Physiological Data
|pdfUrl=https://ceur-ws.org/Vol-2563/aics_40.pdf
|volume=Vol-2563
|authors=P. Nisha,Urja Pawar,Ruairi O'Reilly
|dblpUrl=https://dblp.org/rec/conf/aics/NishaPO19
}}
==Interpretable Machine Learning Models for Assisting Clinicians in the Analysis of
Physiological Data==
https://ceur-ws.org/Vol-2563/aics_40.pdf
Interpretable Machine Learning Models for
Assisting Clinicians in the Analysis of
Physiological Data
P. Nisha, Urja Pawar and Ruairi O’Reilly
Cork Institute of Technology, Ireland,
p.nisha@mycit.ie, urja.pawar@mycit.ie, ruairi.oreilly@cit.ie,
www.cs.cit.ie
Abstract. The analysis of physiological data plays a significant role
in medical diagnostics. While state-of-the-art machine learning models
demonstrate high levels of performance in classifying physiological data
clinicians are slow to adopt them. A contributing factor to the slow rate
of adoption is the “black-box” nature of the underlying model whereby
the clinician is presented with a prediction result, but the rationale for
that result omitted or not presented in an interpretable manner.
This gives rise to the need for interpretable machine learning models
such that clinicians can verify, and rationalise, the predictions made by
a model. If a clinician understands why a model makes a prediction,
they will be more inclined to accept a models assistance in analysing
physiological data.
This paper discusses some of the latest findings in interpretable machine
learning. Thereafter, based on these findings, three models are selected
and implemented to analyse ECG data that are both accurate and exhibit
a high level of interpretability.
Keywords: Interpretable Machine Learning, Decision Trees, Random Forest,
Feature Engineering, ECG, Medical Diagnostics
1 Introduction
In recent years there has been an increased push towards utilising machine learn-
ing as part of healthcare solutions [2] but its adoption has been slow. A major
obstacle to the adoption of machine learning in the clinical decision-making pro-
cess is the black-box nature of the algorithms upon which a machine learning
model may rely upon [3].
A clinician is hesitant to treat a patient’s diagnosis as an input-output process
whereby they feed patient data into a system, and a diagnosis is returned without
any insight as to how that diagnosis was arrived at. Given the significance of the
diagnostic decision-making process, it is understandable that a clinician would
not trust something they do not understand, particularly as patient care needs
Copyright © 2019 for this paper by its authors. Use permitted under Creative Commons License
Attribution 4.0 International (CC BY 4.0).
explanation and the algorithm or models these systems are built upon do not
always provide clarification or a rationale for the prediction made [3].
As such traditional performance metrics (accuracy, precision, f1-Score, recall)
should not be the only consideration when designing a model intended to assist
clinicians in diagnostics, where data can be heterogeneous and where edge cases
cannot be easily identified beforehand. In the clinical world, the stakes are much
higher and as such, the performance of a model should be robust across the
target patient population, its implementation ought to ensure proper use, and
its analysis or predictions should provide context that aids interpretability [6].
In healthcare, interpretability is of utmost importance as the prediction of a
model must be backed by plausible and explainable reasoning as a patients life
stands at risk. However, studies have shown that there is a trade-off between in-
terpretability and performance for machine learning models. Incorporating inter-
pretability as a performance metric for a machine learning could be a significant
step in addressing this problem. Providing clinicians with an insight as to the
suitability of a model to their needs, assisting in predictions and leaving room
for the domain experts to debug and improve the underlying machine learning
models being utilised.
2 Related Work
In the last few years there has been research carried out to come up with a
model that is highly accurate as well as interpretable e.g., GA2M [4], falling
rule lists [11], ensemble decision trees [12] and model distillation [10]. Although
these models have exhibited good performance, they have not yet been used in
the health care industry due to the rarity of their application [3]. A number of
interpretable models are reviewed to form an understanding of the characteristics
that make a model interpretable.
2.1 Decision Trees
A decision tree is a machine learning model that distributes data into subsets and
predicts outcomes based on decision rules(if-then-else rules). The partitioning of
data begins with a binary split and continues until it cannot be split any further.
Different branches of different lengths are formed.
A decision tree captures the training data in the smallest possible tree. This
is done in to simplify the explanation of the problem. Smaller decision trees
provide faster decisions and are easier to understand.
The reasoning approach behind the decision tree readily approachable when
browsing through the tree. This approachability makes the decision trees an
interpretable machine learning model. The attributes that contribute towards
the accuracy and decision making are only included in the rules of the decision
tree. The other attributes are all ignored. This reduced focus helps to provide
useful information about the features.
A decision tree has low bias and high variance. Combining several decision
trees will decrease variance and maintain low bias. The technique of combining
multiple decision trees (models typically) is known as an ensemble method. En-
semble methods perform better than a single decision tree and provide accurate
predictions.
Ensemble Decision Trees Ensemble decision trees can be formed using two
techniques:
Bagging — Bagging is a general technique for combining the predictions of
many models. This technique uses randomly sampled training sets to train
each tree. An ensemble of different models is obtained as a result.
Bootstrap Aggregation — Boosting trains the models iteratively such
that training of any model or tree at the current step depends on the previ-
ously trained model or tree. Each new model or tree gives more importance
to train the observations that were badly handled by a previous model. This
helps in obtaining a strong predictor with low bias.
For a classification problem, the aggregated prediction will be generated by
a majority vote from all the different models or trees. Random Forests are an
extension of bagging. Random features are selected to build a tree instead of
using all the features. The feature that gives the best split is used to split the
node of the tree. A group of random decision trees constitute a random forest.
In [12], an ensemble of decision trees was used to classify ECG data, and
an accuracy of 90.4% was obtained. The ECG signal of frequency 0.5Hz 40Hz
was only considered for the implementation as it was considered as the most
important part of the signal. Apart from this frequency range, everything else
was considered to fall under noise and was eliminated. The ensemble decision
tree was generated using the Bootstrap Aggregation method.
2.2 GA2MS
Generalised Additive Model (GAM) is an additive modelling technique which
can be used to capture nonlinearities in the data. The contribution of each in-
dependent variable to the prediction of GAM model is clearly stated, making it
highly interpretable. GA2MS is an extended form of the GAM model. GA2MS
model is obtained when pairwise interactions (considering the interactions be-
tween two different features) are added to the GAM model.
In paper [4], two case studies were presented where generalised additive mod-
els with pairwise interactions (GA2MS) were applied to healthcare problems, and
state-of-the-art accuracies were obtained.
The GA2MS model in [4] was used to detect the probability of death due to
pneumonia in patients. This was done to ensure the patients with high risk can
be immediately attended. Every feature or term in the model returned a risk
score. A risk score above zero indicated a higher risk, and below zero indicated
a lower risk. All the risk scores for a particular patient were added together
(to a baseline risk). The aggregate risk was converted to a probability score.
Features were selected on the basis of the risk score. Selecting the most critical
features made the patient’s status more understandable (by the clinical expert)
and contributed to the interpretability of the model.
Spline interpolation was used to overcome irregularities in the data. The
model was trained with spline, which considers each point in the data and rep-
resents it to form a smooth curve; this also helps tackle the overfitting of the
model. The GA2MS model detected patterns in the data which were missed be-
fore by other models. The feature selection technique of this model provided an
accuracy of 85.7%. Most importantly, this paper demonstrated how predictions
made by the model could be explained for an individual patient by considering
only the most critical features (depending on the risk score)[4].
2.3 Model Distillation
Model distillation is a technique which makes use of two different machine learn-
ing models, namely - a student model and a teacher model. A teacher model is a
complex machine learning model like neural networks, whereas a student model
is an interpretable model like decision trees. The main aim of model distillation
is to transfer the generalisations or learnings of the complex teacher model to
the interpretable student model. This way, the reasons behind the predictions
made by the complex black-box models can be made easily understandable by
the student model.
The complex teacher model is well trained and regularised (to avoid overfit-
ting) to perform well on unseen data. The training knowledge obtained by the
teacher model is then distilled to the student model.
In this paper [10], a transparent model distillation technique was used to un-
derstand and detect the bias in black-box models. This was achieved by training
a transparent student model to mimic the black-box model and then comparing
the transparent mimic model to a transparent model using the same features
on “true” outcomes instead of the labels predicted by the black-box model.
Difference between the transparent mimic and true-label model portray how a
black-box model predicts and how a model trained on “true” outcomes highlights
potential bias in the black-box model.
Interpretability mainly means transparency of the features used and easily
understandable algorithm, and it differs from person to person and the use case.
In a broader sense, interpretability can be described as transparency of the
machine learning model, i.e., the algorithm, features, parameters and the model
should be comprehensible by the end-user [8].
3 Methodology
Studies have shown that there is a trade-off between interpretability and per-
formance in machine learning models [5]. This means that while models like
decision trees and regression models are highly interpretable, they are less ac-
curate when compared to less interpretable models like neural networks and
other deep learning models. Thus, one has to compromise on either of the two
attributes (interpretability and accuracy).
Generalized Linear Model was selected because it is fast, computationally
inexpensive and interpretable in nature. A Decision Tree model was selected as
it requires very little data preparation and is very intuitive and easily explainable.
Thirdly, Random Forests was selected as it is one of the most accurate learning
algorithms and can handle data imbalance and variance in data implicitly.
The ECG dataset was preprocessed in a way that removed all noise such that
more accurate results could be obtained. After having preprocessed the signals in
the dataset, it was trained and validated on three different interpretable models
Generalized Linear Model, Decision Trees and Random Forests. Since Random
Forest is not interpretable by its nature, Graphviz was used to generate the tree
structure to aid interpretation as to why a prediction was made.
While analysing ECG data, making an accurate prediction only partially
solves the problem. The reason why it is considered as an accurate prediction or
why a certain prediction was made adds value to the analysis.
Section 3.1 contains the details about the dataset used. Section 3.2 discusses
the various pre-processing techniques employed. Section 3.3 discusses the feature
engineering techniques and section 3.4 provides a detailed description of each of
the models used.
3.1 Data
The dataset being used is the Physionet MIT-BIH dataset [7] available from
Kaggle [1] and originally presented in [9]. The dataset is already normalized, and
the R-R interval is extracted by applying a threshold of 0.9 on the normalized
value. Since the signals before normalization were of different morphologies, the
R-R beats that were extracted were padded with zeroes to achieve an equal
length. The R-R beats present in the dataset are all of equal length. There are
5 classes in the dataset and the normal class is heavily oversampled that is, a
disproportionately high sample of normal class is present in the dataset. There
are total of 5 classes with each corresponding a particular heart condition as
denoted in Table 1.
Class Heart condition
N Normal, Left/Right bundle branch block, Atrial escape, Nodal escape
S Atrial premature, Aberrant atrial premature, Nodal premature, Supra-ventricular premature
V Premature ventricular contraction, Ventricular escape
F Fusion of ventricular and normal
Q Paced, Fusion of paced and normal, Unclassifiable
Table 1. The different classes in the MIT-BIH dataset.
While recording an ECG signal, it can be contaminated by a variety of in-
terfering signals that are classified as noise. The source of this noise can be a
patients movement, respiration, surrounding disturbances, muscle movements
etc. Noise degrades the signal quality which leads to misinterpretation. Thus, it
is necessary to de-noise the signal before it can be used in diagnosis which is
discussed in section 3.2.
3.2 Preprocessing
Signal preprocessing is used to eliminate the noise from the signal and is an
important process in increasing the performance of heart beat classification.
The following are the signal preprocessing techniques carried out to the data.
1. Differencing
2. Normalising
3. Smoothing
Differencing – A stationary time series signal is one in which the components
(mean, variance and co-variance) do not vary with time (independent of time).
The underlying assumption in signal preprocessing techniques is that the signal
is stationary. It is simpler to analyse stationary signals as the complexity of the
time component is not taken into account.
Non-stationary time series data can be transformed into stationary time se-
ries data by applying a preprocessing technique called Differencing. Differencing
is carried out by subtracting the previous observation (data point lying in front
in the series) from the current one. It can help stabilize the mean of time-series
data by removing or eliminating the effects of trend and seasonality.
Differencing is carried out by using the Pandas diff() function. The function is
useful as it maintains the date and time information and satisfies the underlying
assumption when processing the signal data. After differencing the data in the
first column it is filled with NaN (Not a number) values as there are no values
to the left of it to find the difference. To address this, Pandas fillna() function is
used to fill in the NaN values, the backward filling method is used on the columns
(i.e. using the next valid data observation to fill the gap in the dataset).
Normalising the signal — A time series signal is normalised to rescale the
data so that all distributions are alike and relevant comparison can be done.
Normalising also reduces noise in the signal. The data was normalized by divid-
ing the column value with the maximum value of the column. This operation
transform the data to the same scale.
Smoothing using Moving Average Function — Smoothing a signal re-
duces the noise. While performing smoothing data points of the signal are ad-
justed. The individual data points that are higher than the immediately adjacent
datapoints (assumed to be due to noise), are lowered. In case the individual data
points are lower than their neighbouring data points, they are increased. This
results in eliminating distortion and a smoother signal is obtained.
The moving average technique was the smoothing technique used, it has
an underlying assumption that independent noise will not change the signal.
According to this assumption, if a few data points are averaged the noise can
be eliminated. Moving average makes use of a window, which is slid across the
whole time series data to calculate the average values. It transforms the old
time series data to a new time series data after averaging the values. The rolling
function available in pandas was used as the moving average function for this
work. The rolling function automatically groups observations into a window
where a window size can be specified, and a trailing window can be created.
For the purpose of this work, the window size of 7 samples was taken. Trailing
window makes use of historical observations and are used for time-series data.
After the trailing window is created it takes the max value and changes the
dataset. Rolling window operations are an important transformation that can
be done on a dataset containing time-series data. The transformed data retains
the same frequency as the original data.
3.3 Feature Engineering
Feature engineering has been used to create a much bigger feature space and
gather more information from the data. More features were manually created
from previously existing ones, to make the predictive model better. It also helped
to increase the model accuracy on unseen data.
The frequency of the time series signal has been changed in this research
by decreasing the frequency of the signal (Downsampling). Downsampling is a
feature engineering technique which helps reduce the signal processing time.
It is done by using scipy signal decimate function which uses the anti-aliasing
filter. The anti-aliasing filter is a low pass filter which only lets low frequencies
pass through and attenuates higher frequencies. The dataset had a sampling
frequency of 125Hz, it was downsampled by a factor of 5 such that the first
element and every fifth element then onwards was persisted.
3.4 Models
Generalised Linear Model A linear regression model predicts the target
as the weighted aggregate of all the input features. Logistic regression is used
in classification problems and it is an extension of linear regression where the
model predicts probabilities based on two possible outcomes. GLM (Generalised
Linear Model) is an extension to linear regression without the assumption that
the outcome distributions will be gaussian in nature.
GLM calculates the expected mean of the non-gaussian outcome distribution
and connects it to the weighted sum of input features by passing it through
a non-linear function. GLM can be defined as a more flexible model by keep-
ing the interpretability intact. Modelling based on the weighted sum makes the
model transparent and provides an explanation as to why certain predictions are
made. Not only predictions on weights, but confidence intervals of the weights
themselves can be derived from analysing the contributing features. The given
problem is a multiclass problem and logistic regression is used for classification.
GLM has been modelled using Logistic Regression.
A solver helps to fit or train the data. Below are the different types of solvers
available in Logistic Regression:
1. Newton-CG solver: The Newton Solver has a very fast converge range (learns
much faster). It uses the principle of gradient descent with the Hessian (a
squared matrix of second order partial derivatives) to achieve faster conver-
gence.
2. Limited-memory-Broyden-Fletcher-Goldfarb algorithm (LBFGS solver): This
solver is similar to a newton solver. The only difference being it uses an es-
timation to the inverse hessian matrix. This saves significant memory but a
major disadvantage is that in some cases it may not converge to anything.
3. A Library for Large Linear Classification or (LibLinear solver): This solver
is a linear classifier that makes decisions based on the linear combination of
features. It performs approximate minimizations along the co-ordinate di-
rections. The main drawback of this solver is that it does not perform well
for multi-class problems.
4. Stochastic Average Gradient or (SAG solver): SAG solvers are best suited
for large datasets with large number of features. Its memory cost is too high
making it impractical most of the time.
5. SAGA solver: Saga solver is a variant of SAG that is suitable for very large
datasets.
The multi-class parameter has two variants in logistic regression:
1. Multinomial: It is a classification method used for data when data has nom-
inal or categorical dependant variables.
2. One vs Rest Approach (Ovr): One vs rest approach can be used to convert
any problem into binary classification problem. This method trains different
distinct binary classifiers, each classifier is designed to predict or recognize
a particular class.
The different Class weights parameters available in logistic regression are:
1. Balanced: It ensures that there is a balance mix of classes by weighing the
classes inversely proportional to their frequency.
2. None: If the class weight is specified as None then the class weights will be
uniform in nature.
The parameter inverse of regularization strength is named as C. Regular-
ization prevents over-fitting of the model. This parameter has been used in the
research and has been assigned a smaller value to specify better regularization.
The smaller the value of C, the stronger the regularization ensuring the model
does not overfit. The dataset is heavily imbalanced and since imbalance was
not addressed during the preprocessing of the data, the balanced class weight
method has been used to handle this.
The selection of the parameters for the Logistic Regression model used are
Newton-cg solver, OVR, C and Balanced.
Decision Trees learn through if-then-else decision rules making the out-
come of the model interpretable and the root-cause of a prediction easy to fol-
low. Decision tree models split the data depending on certain cut-off values in
their features. Different subsets of the dataset are created when the nodes are
split and an associated tree gets generated incrementally. Finally, a tree with leaf
nodes and decision nodes is obtained. A decision node has two or more branches.
A leaf node represents the classification or decision. The top most decision
node is called the root node and is considered as the best predictor. The inter-
pretation of the decision tree is also very simple. The root node is the starting
point and the next nodes classify all the subsets. Once the leaf node is reached
the predicted output can be obtained.
The simplicity of the interpretation is often contributed to as data ends up in
distinct groups making it easier to understand. Sklearn’s decision tree classifier is
used to model a Decision tree. The export graphviz function is used to visualize
the decision tree. The graphviz function provides a detailed graph with the tree’s
structure containing the if-then-rules. This helps to understand why a certain
decision was taken.
Random Forest Random forest is a very flexible machine learning model. It
creates a forest, which is an ensemble of decision trees, trained with the bagging
method. Bagging employs the idea that the combination of learning models
increases the overall results. Random Forests generates multiple decision trees
and merges them together to get a more accurate and stable prediction.
The importance of each feature on the prediction or outcome can be found
out using sklearns feature importance. This technique is used to analyse each
features importance and to predict what led to the outcome of the model. Thus,
knowing the contribution of each feature turns this into a white-box model.
4 Results
Performance Measures: The performance of the classification algorithms were
evaluated using five measures:
Confusion Matrix: Terms associated — True Positives (TP), True Nega-
tives (TN), False Positives (FP) and False Negatives (FN)
Precision — Precision of a model can be defined as the ratio of correctly
predicted positive observations to the total predicted positive observations.
Precision = TP/TP+FP
Recall — Recall of a model can be defined as the ratio of correctly predicted
positive observations to the all observations in actual true class. Recall =
TP/TP+FN
Classification accuracy — Accuracy of a model can be defined as a ratio
of correctly predicted observation to the total number of observations made.
Accuracy = TP+TN/TP+FP+FN+TN
F1 score — F1 score of a model can defined as the weighted average of Pre-
cision and Recall. F1 Score = 2*(Recall * Precision) / (Recall + Precision)
GLM Random Forest Decision Tree
Classes Precision Recall F1 Precision Recall F1 Precision Recall F1
N 0.98 0.89 0.93 0.97 1.00 0.98 0.97 0.98 0.98
S 0.31 0.65 0.42 0.99 0.60 0.75 0.65 0.65 0.65
V 0.69 0.83 0.75 0.97 0.84 0.90 0.86 0.84 0.85
F 0.15 0.78 0.25 0.87 0.56 0.68 0.52 0.60 0.56
Q 0.89 0.94 0.91 0.99 0.93 0.96 0.94 0.94 0.94
Average 0.93 0.88 0.90 0.97 0.97 0.97 0.95 0.95 0.95
Table 2. Table depicting performance results for the three models implemented: Gen-
eralized Linear Model (GLM), the Random Forest and the Decision Tree.
Model Accuracy
GLM 88.29%
Random Forest 97%
Decision Tree 95%
Table 3. Accuracy scores of the three models evaluated.
Table 2 depicts the performance metrics of GLM model. It can be seen that
the model has a high precision average of 0.93 and a recall of 0.88. The accuracy
obtained by the model is 88.29%. The Random Forest model also has a high
precision average of 0.97 and a recall of 0.97. The accuracy obtained by the
model is 97.00% and the model could classify almost all of the classes correctly.
Finally, the Decision Tree model was shown to have a high precision average of
0.95 and a recall of 0.95. The accuracy obtained by the model is 95.00%.
Table 3 denotes the accuracy scores of all the three models. Random Forest
performed the best out of all the models obtaining an accuracy 97%. The model
has a very high precision rate for all the classes as can be seen from the table.
4.1 Hyper Parameter Tuning of Generalized Linear Model
The Generalized Linear Model was evaluated on a variety of parameters such
as solver, multi-class, regularization factor(c) and class-weight. The model was
tested on four different solvers. The GLM model performed the best using the fol-
lowing parameters: (multi class = ovr, solver = newton-cg, class weight
= balanced, c = 0.5). Table 4 denotes the accuracy obtained for the solvers
evaluated.
Solver Accuracy
SAG 73.41%
SAGA 82.40%
Newton-cg 88.29%
LBFGS 87.23%
Table 4. Results of the different solvers evaluated for the GLM model.
5 Conclusion
In areas such as medical diagnostics, accuracy is not the only factor that de-
termines the performance of a machine learning model. Interpretability plays a
crucial role due to its importance in understanding the rationales for a model’s
predictions. As such, interpretability is particularly crucial in models designed
for the analysis of physiological data and clinical purposes.
In this paper, MIT-BIH dataset is used for heartbeat classification. Prior
to modelling of the data, various preprocessing techniques have been used to
eliminate noise in the ECG signal. The noise present in the ECG signal leads
to misinterpretation. Feature engineering techniques have been employed to im-
prove the performance of the models.
The data has then been trained on three different models namely - Gener-
alized Linear Model, Random Forests and Decision Trees. Graphviz has been
used to convert the black box Random Forest model to an interpretable model.
All the three models provide comprehensible decisions of the predictions made.
An accuracy of 97% on using Random Forest model was obtained. The obtained
accuracy is comparable to state of the art models in ECG classification.
The models presented achieved a high level of accuracy as well as a high level
of interpretability. The ability to employ techniques such as feature importance
that identify the underlying features contributing towards the decision taken by
a model increases the transparency in the classification process, making it more
akin to a white-box model. Providing clinicians with an overview of features, and
associated values, contributing to a decision enable the clinician to deduce the
rationale behind a prediction. It is envisaged that this would increase clinicians’
trust and confidence in a prediction and assist them in providing excellent patient
care.
Acknowledgement: This material is based upon works supported by Science
Foundation Ireland under Grant No. SFI CRT 18/CRT/6222
References
1. Mit-bih arrhythmia database. https://www.kaggle.com/mondejar/
mitbih-database
2. Ahmad, M.A., Eckert, C., Teredesai, A.: Interpretable machine learning in health-
care. In: Proceedings of the 2018 ACM International Conference on Bioinformatics,
Computational Biology, and Health Informatics. pp. 559–560. ACM (2018)
3. Ahmad, M.A., Eckert, C., Teredesai, A.: Interpretable machine learning in health-
care. In: Proceedings of the 2018 ACM International Conference on Bioinformatics,
Computational Biology, and Health Informatics. pp. 559–560. ACM (2018)
4. Caruana, R., Lou, Y., Gehrke, J., Koch, P., Sturm, M., Elhadad, N.: Intelligible
models for healthcare: Predicting pneumonia risk and hospital 30-day readmission.
In: Proceedings of the 21th ACM SIGKDD International Conference on Knowledge
Discovery and Data Mining. pp. 1721–1730. ACM (2015)
5. Doshi-Velez, F., Kortz, M., Budish, R., Bavitz, C., Gershman, S., O’Brien, D.,
Schieber, S., Waldo, J., Weinberger, D., Wood, A.: Accountability of ai under the
law: The role of explanation. arXiv preprint arXiv:1711.01134 (2017)
6. Glass, A., McGuinness, D.L., Wolverton, M.: Toward establishing trust in adaptive
agents. In: Proceedings of the 13th international conference on Intelligent user
interfaces. pp. 227–236. ACM (2008)
7. Goldberger AL, Amaral LAN, G.L.H.J.I.P.M.R.M.J.M.G.P.C.K.S.H.: Physiobank,
physiotoolkit, and physionet: Components of a new research resource for complex
physiologic signals. IEEE Engineering in Medicine and Biology Magazine 101(23),
215–220 (2003)
8. Lipton, Z.C.: The mythos of model interpretability. arXiv preprint
arXiv:1606.03490 (2016)
9. Moody, G.B., Mark, R.G.: The impact of the mit-bih arrhythmia database. IEEE
Engineering in Medicine and Biology Magazine 20(3), 45–50 (2001)
10. Tan, S., Caruana, R., Hooker, G., Lou, Y.: Detecting bias in black-box models
using transparent model distillation. arXiv preprint arXiv:1710.06169 (2017)
11. Wang, F., Rudin, C.: Falling rule lists. In: Artificial Intelligence and Statistics. pp.
1013–1022 (2015)
12. Zaunseder, S., Huhle, R., Malberg, H.: Cinc challengeassessing the usability of ecg
by ensemble decision trees. In: 2011 Computing in Cardiology. pp. 277–280. IEEE
(2011)