=Paper=
{{Paper
|id=Vol-2283/MediaEval_18_paper_55
|storemode=property
|title=Weighted Discriminant Embedding: Discriminant Subspace Learning for Imbalanced Medical Data Classification
|pdfUrl=https://ceur-ws.org/Vol-2283/MediaEval_18_paper_55.pdf
|volume=Vol-2283
|authors=Tobey H. Ko,Zhonglei Gu,Yang Liu
|dblpUrl=https://dblp.org/rec/conf/mediaeval/KoGL18
}}
==Weighted Discriminant Embedding: Discriminant Subspace Learning for Imbalanced Medical Data Classification==
https://ceur-ws.org/Vol-2283/MediaEval_18_paper_55.pdf
Weighted Discriminant Embedding: Discriminant Subspace
Learning for Imbalanced Medical Data Classification
Tobey H. Ko1 , Zhonglei Gu2 , Yang Liu2,3
1
Department of Industrial and Manufacturing Systems Engineering, University of Hong Kong, HKSAR, China
2
Department of Computer Science, Hong Kong Baptist University, HKSAR, China
3
Institute of Research and Continuing Education, Hong Kong Baptist University, Shenzhen, China
tobeyko@hku.hk,csygliu@comp.hkbu.edu.hk,cszlgu@comp.hkbu.edu.hk
ABSTRACT W β Rπ·Γπ (π β€ π·), which is capable of projecting the o-
A model designed for automatic prediction of diseases based riginal high-dimensional data to a low-dimensional subspace
on multimedia data collected in hospitals is introduced in π΅ = Rπ , where the weighted discriminant information could
this working notes paper. In order to perform the automatic be preserved.
diseases prediction efficiently, while using as few data as pos- In this yearβs Medico task, the sample numbers in different
sible for training, we develop a two-stage learning strategy, classes are highly imbalanced. To enhance the algorithmβs
which first performs the weighted discriminant embedding power in making correct detection on rarer classes, we expect
(WDE) to project the original data to a low-dimensional that data samples belonging to the same class, especially for
feature subspace and then utilizes the cost-sensitive nearest the rarer class, should be close to each other as much as pos-
neighbor (CS-NN) method in the learned subspace for dis- sible in the learned subspace, while nearby data samples from
ease prediction. The proposed approach is evaluated on the different classes, again, especially for rarer classes, should be
MediaEval 2018 Medico Multimedia Task. separated from each other as much as possible in the learned
subspace.
To minimize the weighted intra-class scatter, we present
1 INTRODUCTION the following objective:
Aiming at improving the efficiency of detecting medical abnor- π
(οΈ βοΈ )οΈ
malities in the machine intelligence assisted medical diagnosis, W = arg min π‘π π΄ππ Wπ (xπ β xπ )(xπ β xπ )π W , (1)
and using as little information as possible, the MediaEval W π,π=1
2018 Medico Multimedia Task [3] seeks to design an integrat- where π΄ππ = (πΌπ + πΌπ )/2 if ππ = ππ ; and 0 otherwise. Here πΌπ
ed approach to assist the medical expertsβ decision-making indicates the importance of class ππ and is defined using the
process using a combination of video and image information, entropy-based formulation [2]:
as well as other sensory information. In this paper, a two-stage
learning strategy is introduced to facilitate efficient detec- (1 β ππ )2
πΌπ = β log(ππ ), (2)
tion of diseases using multimedia and sensory information. ππ
The first stage consists of a dimensionality reduction process where ππ denotes the proportion of class ππ in the dataset. In
which projects the original data to a low-dimensional fea- Eq. (2), small proportion indicates high importance. Eq. (1)
ture representation using weighted discriminant embedding could be rewritten as:
(WDE), which improves the efficiency of the learning process
W = arg min π‘π(Wπ Lπ΄ W), (3)
while also preserving the key discriminant information of W
the original data. Then, the cost-sensitive nearest neighbor
where Lπ΄ is a Laplacian matrix [1] defined as Lπ΄ = Dπ΄ β
(CS-NN) method is employed to make the prediction in the
βοΈπwith Dπ΄ being a diagonal matrix defined as (π·π΄ )ππ =
A,
learned subspace.
π=1 (π΄)ππ (π = 1, Β· Β· Β· , π).
Similarly, we define the following objective function to
2 WEIGHTED DISCRIMINANT
maximize the weighted inter-class scatter:
EMBEDDING (οΈ βοΈ π )οΈ
Let π³ be the training set: π³ = {(x1 , π1 ), Β· Β· Β· , (xπ , ππ )}, where W = arg max π‘π π΅ππ Wπ (xπ β xπ )(xπ β xπ )π W , (4)
xπ β Rπ· (π = 1, ..., π) denotes the feature representation of W π,π=1
the π-th sample, ππ β {1, Β· Β· Β· , πΆ} denotes the label of xπ , π where π΅ππ = πππ (πΌπ + πΌπ )/2 if ππ ΜΈ= ππ ; and 0 otherwise. Here
denotes the number of data samples in the set, πΆ denotes πππ = ππ₯π(ββxπ β xπ β2 /2π 2 ) is utilized to measure the close-
the number of classes, and π· denotes the original dimen- ness between two data samples. Eq. (4) could be rewritten
sion of data. Given the training set, weighted discriminant as:
embedding (WDE) aims to learn a transformation matrix
W = arg max π‘π(Wπ Lπ΅ W), (5)
W
Copyright held by the owner/author(s).
MediaEvalβ18, 29-31 October 2018, Sophia Antipolis, France where Lπ΅ = Dπ΅ β βοΈ B, with Dπ΅ being a diagonal matrix
defined as (π·π΅ )ππ = π
π=1 (π΅)ππ (π = 1, Β· Β· Β· , π).
MediaEvalβ18, 29-31 October 2018, Sophia Antipolis, France T. H. Ko, Z. Gu, Y. Liu
We integrate Eqs. (3) and (5) to form a unified objective Table 1: Results of our approach on the first subtask
function of WDE: of MediaEval 2018 Medico Multimedia Task.
(οΈ π )οΈ
W Lπ΅ W Recall Precision Accuracy F1 Score Rk
W = arg max π‘π . (6)
W Wπ Lπ΄ W
Run 1 0.5001 0.4917 0.9471 0.4830 0.5357
Then the optimal W that maximizes the objective func- Run 2 0.4415 0.4294 0.9384 0.4251 0.4612
tion in Eq. (6) is composed of the normalized eigenvectors Run 3 0.3947 0.3670 0.9320 0.3728 0.4035
corresponding to the π largest eigenvalues of the following Run 4 0.3553 0.3333 0.9256 0.3324 0.3511
eigen-decomposition problem: Run 5 0.3019 0.2814 0.9186 0.2812 0.2918
Lπ΅ w = πLπ΄ w. (7)
For a high-dimensional data sample xπ , it can be mapped to Table 2: Results of our approach on the second sub-
the subspace by yπ = Wπ xπ . task of MediaEval 2018 Medico Multimedia Task.
3 RESULTS AND ANALYSIS Recall Precision Accuracy F1 Score Rk
To evaluate our approach, we test its performance on the Run 1 0.5005 0.4917 0.9471 0.4830 0.5357
MediaEval 2018 Medico Multimedia Task. The task contains Run 2 0.4181 0.3857 0.9337 0.4251 0.4193
both development set (with 5, 293 samples) and test set (with Run 3 0.4259 0.4085 0.9350 0.4040 0.4348
8, 740 samples). For each sample, we use six types of features: Run 4 0.3430 0.3107 0.9231 0.3135 0.3293
the 168-D JCD feature; the 18-D Tamura feature; the 33-D Run 5 0.3257 0.3053 0.9227 0.3057 0.3246
ColorLayout feature; the 80-D EdgeHistogram feature; the
256-D AutoColorCorrelogram feature; and the 630-D PHOG
feature. The totally dimension is 1, 185. The reason might be that the proposed WDE is a linear map-
We participate in two subtasks: 1) Classification of diseases ping method, which is not sufficient to capture the complex
and findings; and 2) Fast and efficient classification. For both discriminant information embedded in the high-dimensional
tasks, we submit 5 runs. feature space. This motivates us to consider extending our
β For Run 1 (on both subtasks), we use all the data method to the nonlinear case to improve the performance.
from the development set for training; Furthermore, by comparing the performance on Run 2 (Run
β For Run 2 (on both subtasks), we randomly select 4) and that on Run 3 (Run 5), we observe that even we use all
50% data for each class from the development set the data from the minority classes (i.e., the βout-of-patientβ
for training; and βinstrumentsβ classes), the performance is not improved.
β For Run 3 (on both subtasks), we randomly select The reason might be that the number of data in these two
50% data for each class from the development set, classes are too small to represent the βrealβ distribution of the
together with the remaining data in the βout-of- classes. On possible solution is to employ the oversampling
patientβ and βinstrumentsβ classes, for training; technology to reasonably and faithfully generate samples for
β For Run 4 (on both subtasks), we randomly select minority classes.
25% data for each class from the development set for
training; 4 CONCLUSION
β For Run 5 (on both subtasks), we randomly select In this paper, we propose a subspace learning method called
25% data for each class from the development set, weighted discriminant embedding (WDE), aiming at discov-
together with the remaining data in the βout-of- ering the discriminant subspace for imbalanced dataset. After
patientβ and βinstrumentsβ classes, for training. dimensionality reduction, the cost-sensitive nearest neighbor
In the training stage, we use the training data to learn is utilized for classification. We plan to extend our work
the transformation matrix W via WDE. We set π = 1 and from two aspects. First, we will generalize our approach to
the subspace dimension π = 50. In the test stage, we use the nonlinear case to enhance its data representation ability. Sec-
obtained W to map both training and test data to the 50-D ond, we will incorporate some oversampling methods into
subspace, and then use the cost-sensitive nearest neighbor our approach to make it stronger for imbalanced learning
(CS-NN) method for the final classification in the learned problem.
subspace, where the cost of misclassifying the data of class
π (π = 1, Β· Β· Β· , πΆ) to other classes is defined as πππ π‘π = π/ππ , ACKNOWLEDGMENTS
with π and ππ being the total number of the training data This work was supported in part by the National Natural Sci-
and the number of data in class π, respectively. ence Foundation of China (NSFC) under Grant 61503317, in
Tables 1 and 2 report the results of our approach on sub- part by the General Research Fund (GRF) from the Research
task 1 and subtask 2, respectively. Although the accuracy Grant Council (RGC) of Hong Kong SAR under Project
looks good, the overall performance is far from satisfactory as HKBU12202417, and in part by the SZSTI Grant with the
the results on other four important criteria are relatively low. Projct Code JCYJ20170307161544087.
Weighted Discriminant Embedding MediaEvalβ18, 29-31 October 2018, Sophia Antipolis, France
REFERENCES
[1] M. Belkin and P. Niyogi. 2003. Laplacian Eigenmaps for
Dimensionality Reduction and Data Representation. Neural
Comput. 15, 6 (2003), 1373β1396.
[2] T. Lin, P. Goyal, R. Girshick, K. He, and P. Dollr. 2017. Focal
Loss for Dense Object Detection. In 2017 IEEE International
Conference on Computer Vision (ICCV). 2999β3007.
[3] K. Pogorelov, M. Riegler, P. Halvorsen, T. de Lange, K. R.
Randel, D.-T. Dang-Nguyen, M. Lux, and O. Ostroukhova.
Medico Multimedia Task at MediaEval 2018. In Proceedings of
the MediaEval 2018 Workshop. CEUR-WS, Sophia Antipolis,
France, 29β31 October, 2018.