=Paper=
{{Paper
|id=None
|storemode=property
|title=Comparing Data Mining with Ensemble Classification of Breast Cancer Masses in Digital Mammograms
|pdfUrl=https://ceur-ws.org/Vol-941/aih2012_GhassemPour.pdf
|volume=Vol-941
}}
==Comparing Data Mining with Ensemble Classification of Breast Cancer Masses in Digital Mammograms==
https://ceur-ws.org/Vol-941/aih2012_GhassemPour.pdf
AIH 2012
Comparing Data Mining with Ensemble
Classification of Breast Cancer Masses in Digital
Mammograms
Shima Ghassem Pour1 , Peter Mc Leod2 , Brijesh Verma2 , and Anthony Maeder1
1
School of Computing, Engineering and Mathematics,
University of Western Sydney
Campbelltown, New South Wales, Australia
2
School of Information and Communication Technology,
Central Queensland University
Rockhampton, Queensland, Australia
{shima.ghassempour,mcleod.ptr}@gamil.com,
b.verma@cqu.edu.au,a.maeder@uws.edu.au
Abstract. Medical diagnosis sometimes involves detecting subtle indi-
cations of a disease or condition amongst a background of diverse healthy
individuals. The amount of information that is available for discover-
ing such indications for mammography is large and has been growing
at an exponential rate, due to population wide screening programmes.
In order to analyse this information data mining techniques have been
utilised by various researchers. A question that arises is: do flexible data
mining techniques have comparable accuracy to dedicated classification
techniques for medical diagnostic processes? This research compares a
model-based data mining technique with a neural network classification
technique and the improvements possible using an ensemble approach. A
publicly available breast cancer benchmark database is used to determine
the utility of the techniques and compare the accuracies obtained.
Keywords: latent class analysis, digital mammography, breast cancer,
clustering, classification, neural network.
1 Introduction
Medical diagnosis is an active area of pattern recognition with different tech-
niques being employed [17, 19, 12]. The expansion of digital information for dif-
ferent cohorts [15] has allowed researchers to examine relationships that were
previously not uncovered due to the limited nature of information as well as a
lack of techniques being available for the analysis of large data sets. Flexible
data mining techniques have the capacity to predict disease and reveal previous
unknown trends.
The question that arises is whether the relationships that are revealed by
those techniques are as accurate or as comparable as techniques that are specif-
ically developed for other purposes, such as a diagnostic system for a particular
55
AIH 2012
2 Comparing Data Mining with Ensemble Classification
disease or condition. This research aims at contrasting the cluster analysis tech-
nique (Latent Class Analysis) of Ghassem Pour, Maeder and Jorm [4] against a
baseline neural network classifier, and then considers the effects of applying an
ensemble technique to improve the accuracies obtained.
The organisation of this paper is that section two provides a background on
the approaches that have been utilised for breast cancer diagnosis, sections three
and four detail the proposed techniques for comparison, section five outlines the
experimental results obtained and conclusions are presented in section six.
2 Background
Medical diagnosis is a problematic paradigm in that complex relationships can
exist in the diagnostic features that are utilised to map to a resultant diagnosis
about the disease state. In different cases the state of the disease condition itself
can be marked by stages where the diagnostic symptoms or signs can be subtle
or different to other stages of the disease. This means that there is often not a
clean mapping between the diagnostic features and the diagnosis [13, 14].
Breast cancer screening using mammography provides an exemplar of this
situation. Early detection and treatment have been the most effective way of
reducing mortality [2] however Christoyianni et al. [1] noted that 10-30% of
breast cancers remain undetected while 15-30% of biopsies are cancerous. Tay-
lor and Potts [22] made similar observations in their research. There are many
reasons why various cancers can remain undetected. These include the obfus-
cation of anomalies by surrounding breast tissue, the asymmetry of the breast,
prior surgery, natural differences in breast appearance on mammograms, the low
contrast nature of the mammogram itself, distortion from the mammographic
process and even talc or powder on the outside of the breast making it hard to
identify and discriminate anomalies. Even if an anomaly is detected, a high rate
of false positives exist [17, 18].
Clustering has provided a widely used mechanism for organising data into
similar groupings. The usage of clustering has also been extended to classifiers
and detection systems in order to improve detection and provide greater classi-
fication accuracy. Kim et al. [9] developed a classifier based on Adaptive Res-
onance Theory (ART2) where micro-calcifications were grouped into different
classes with a three-layered back propagation network performing the classifica-
tion. The system achieved 90% sensitivity (Az of 0.997) with a low false positive
rate of 0.67 per cropped image.
Other researchers such as Mohanty, Senapati and Lenka [16] explored the
application of data mining techniques to breast cancer diagnosis. They indi-
cated that data mining medical images would allow for the collection of effective
models, rules as well as patterns and reveal abnormalities from large datasets.
Their approach was to use a hybrid feature selection technique with a decision
tree classifier to classify breast cancer. They utilised 300 images from the MIAS
database. They achieved a classification accuracy of 97.7% however their dataset
images contained microcalcifications as well as mass anomalies.
56
AIH 2012
Comparing Data Mining with Ensemble Classification 3
3 Latent Class Analysis and Data Mining
Latent Class Analysis (LCA) has been proposed as a mechanism for improved
clustering of data over traditional clustering algorithms like k-means [11]. LCA
classifies subjects into one of K unobserved classes based on the observed data,
where K is a constant and known parameter. These latent or potential classes
are then refined based upon their statistical relationships with the observed vari-
ables.
LCA is a probabilistic clustering approach: although each object is assumed
to belong to one cluster, there is uncertainty about an object’s membership of
a cluster [11, 10]. This type of approach offers some advantages in dealing with
noisy data or data with complex relationships between variables, although as an
iterative method there is always some chance that it will be susceptible to noise
and in some cases fail to converge.
An advantage of using a statistical model is that the choice of the clus-
ter criterion is less arbitrary. Nevertheless, the log-likelihood functions corre-
sponding to LC cluster models may be similar to the criteria used by certain
non-hierarchical cluster techniques [18]. Another advantage of the model-based
clustering approach is that no decisions have to be made about the scaling of the
observed variables: for instance, when working with normal distributions with
unknown variances, the results will be the same irrespective of whether the vari-
ables are normalized or not.
Other advantages are that it is relatively easy to deal with variables of mixed
measurement levels (different scale types) and that there are more formal cri-
teria to make decisions about the number of clusters and other model features
[3]. We have successfully applied LCA for cases in health data mining when the
anomalous range of variables results in more clusters than have been expected
from a causal or hypothesis based approach [5]. This implies that in some cases
LCA may be used to reveal associations between variables that are more subtle
and complex.
Unsupervised clustering requires prior specification of the number of clusters
K to be constructed, implying that a model for the data is necessary which pro-
vides K. The binary nature of the diagnosis problem implies that K=2 should
be used in ideal circumstances, but the possibility exists that allowing more
clusters would give a better solution (e.g. by allowing several different classes
within the positive or negative groups). Consequently a figure of merit is needed
to establish that the chosen K value is optimal. In this research the Bayesian
Information Criteria (BIC) is determined for the mass dataset in order to gauge
the best number of clusters.
Repeated application of the clustering approach can also lead to different so-
lutions due to randomness in starting conditions. In this work we used multiple
applications of the clustering calculations to allow improvement in the results,
in an ensemble-like approach. Our improvement strategy was based on selection
of the most frequent membership of classes per element, over different numbers
of clustering repetitions.
57
AIH 2012
4 Comparing Data Mining with Ensemble Classification
4 Neural Network and Ensemble Methods
Neural networks have been advocated for breast cancer detection by many re-
searchers. Various efforts to refine classification performance have been made,
using a number of strategies involving some means of choice between alternatives.
Ensembles have been proposed as a mechanism for improving the classification
accuracy of existing classifiers [6] providing that constituents are diverse.
Zhang et al. [23] partitioned their mass dataset obtained from the DDSM
into several subsets based on mass shape and age. Several classifiers were then
tested and the best performing classifier on each subset was chosen. They used
SVM, k-nearest neighbour and Decision Tree (DT) classifiers in their ensemble
and achieved a combined classification accuracy of 72% that was better than
any individual classifier.
Surrendiran and Vadivel [21] proposed a technique that could determine what
features had the most appropriate correlation on classification accuracy and
achieved 87.3% classification accuracy. They achieved this by using ANOVA
DA, Principal Component Analysis and Stepwise ANOVA analysis to determine
the relationship between input feature and classification accuracy.
Mc Leod and Verma [14] utilised a clustered ensemble technique that relied
on the notion that some patterns could be readily identified through cluster-
ing (atomic). Other patterns that were not so easily separable (non-atomic)
were classified by a neural network. The classification process involved an initial
lookup to determine if a pattern was associated with an atomic class however
for non-atomic classes a neural network ensemble that had been created through
an iterative clustering mechanism (to introduce diversity into the ensemble) was
employed. The advantage of this technique is that the ensemble was not ad-
versely affected by outliers (atomic clusters). This technique was applied to the
same mass dataset as utilised in this research and achieved a classification accu-
racy of 91%.
The ensemble utilised in this research was created by fusing together (using
the majority vote algorithm) constituent neural networks that were created by
varying the number of neurons in the hidden layer to create diverse networks for
incorporation into an ensemble classifier.
5 Experimental Results
The experiments were conducted for LCA and neural network techniques and
the related ensemble approaches using mass type anomalies from the Digital
Database of Screening Mammography (DDSM) [7]. The features used for classi-
fication purposes coincided with the Breast Imaging Reporting and Data System
(BI-RADS) as this is how radiologists classify breast cancer. The BI-RADS fea-
tures of density, mass shape, mass margin and abnormality assessment rank are
used as they have been proven to provide good classification accuracy [20]. These
features are then combined with patient age and a subtlety value [7].
Experiments were performed utilising the clustering technique of Ghassem
58
AIH 2012
Comparing Data Mining with Ensemble Classification 5
Pour, Maeder and Jorm [4] on this dataset. This was achieved using the La-
tent Gold R software package. The first step was to utilise the analysis feature of
LatentGold R to calculate the BIC value and the classification error rate. This
information appears in Table 1 below, with Npar designating the resulting pa-
rameter value associated with the LCA.
Table 1. LCA Cluster optimisation based on Classification Error.
Clusters BIC Npar Classification Error
2 1238.8 30 0.0303
3 1240.6 38 0.0403
4 1241.8 46 0.0446
5 1254.1 54 0.0470
Minimisation of BIC and the Classification Error determines the best number
of clusters for the LCA analysis in terms of classification accuracy and this was
found to be 2 clusters. Nevertheless, it might be expected that some further
complexity could be identified in higher numbers of clusters, where multiple
clusters may exist for either positive or negative classes. The results obtained
when cases of more than 2 clusters were merged to form the dominant positive
and negative classes, are detailed in Table 2. These results show the instability
Table 2. LCA Classification Technique Accuracy.
Clusters Accuracy %
2 87.2
3 56.7
4 43.2
5 32.8
of LCA classification for this dataset at higher numbers of clusters, for example
the 2-cluster solution gives better accuracy than the 3-cluster solution (merging
into 2 clusters) and so forth. From this we conclude that the natural 2-cluster
solution is indeed optimal.
In order to provide a comparison, further experiments were performed using
a neural network and then applying an ensemble classifier. The neural network
and ensemble techniques were implemented in MATLAB R utilising the neural
network toolbox. The parameters utilised are detailed in the Table 3 below.
Experiments were first performed with a neural network classifier alone, in order
to provide a baseline for measuring the classification accuracy on the selected
dataset. The results obtained are detailed in Table 4 below. Further experiments
were then performed utilising an ensemble technique with a summary of the
neural network test results using ten-fold cross validation, as detailed in Table
5 below.
59
AIH 2012
6 Comparing Data Mining with Ensemble Classification
Table 3. Neural network configuration parameters.
Parameter Value
Hidden Layers 1
Transfer Function Tansig
Learning Rate 0.05
Momentum 0.7
Maximum Epochs 3000
Root Mean Square Goal 0.001
Table 4. Neural network classification technique accuracy.
Hidden Neurons Accuracy (%)
13 80
25 80
52 90
111 79
Table 5. NN-ensemble classification technique accuracy.
Networks Hidden Neurons in Ensemble Accuracy (%)
6 24,5,15,32,31,43 94
10 24,5,15,32,31,43,50,75,38,59 96.5
13 24,5,15,32,31,43,50,75,38,59,68,79,116 98
15 24,5,15,32,31,43,50,75,38,59,68,79,116,146,14 96
Experiments were also performed for the ensemble-like optimising of results
from the LCA technique. It is difficult to match this process directly with the
complexity used for the NN-ensemble experiments, so the number of repetitions
has been modelled on plausible choice based on dataset size of 100 cases. The
results for these experiments are shown in Table 6 below.
Table 6. LCA-ensemble classification technique accuracy.
Repetitions Accuracy (%)
10 87
20 89
40 91
70 94
6 Discussion and Conclusions
Examination of the results from Tables 1 to 6 demonstrates that the accuracy
obtained with the LCA technique is below that of the baseline classification
60
AIH 2012
Comparing Data Mining with Ensemble Classification 7
performed with the neural network. However an ensemble oriented approach en-
abled improvement of the results from both techniques.
In order to examine the results more closely the sensitivity, specificity and
positive predictive value have been calculated for the best performing results for
each of the trialled techniques, shown below in Table 7.
Sensitivity is the True Positive diagnosis divided by the True Positive and
False Negative components. Sensitivity can be thought of as the probability of
detecting cancer when it exists.
Specificity is the True Negative component divided by the True Negative
component plus the False Positive component. Specificity can be thought of as
the probability of being correctly diagnosed as not having cancer.
Positive Predictive Value (PPV) is the True Positive component divided by
the True Positive component plus the False Positive component. PPV is the accu-
racy of being able to identify malignant abnormalities. The latent class analysis
Table 7. Performance results for the proposed techniques.
Technique Performance(%)
Sensitivity Specificity PPV
Latent Class Analysis 80.5 93.9 95.0
LCA-ensemble 82.7 95.2 96.0
Neural Network 91.6 88.4 90.0
NN-ensemble 97.0 97.9 99.0
technique was not as sensitive as the neural network but had better specificity
and a higher positive predictive value than the neural network. Both ensemble
approaches resulted in substantially better performance, which of course must
be traded off against the increased computational cost. The NN-ensemble tech-
nique performed the best with good sensitivity, specificity and a high positive
predictive value.
The flexibility of clustering techniques such as LCA provides a mechanism for
gaining insight from large data repositories. However once patterns in the data
become evident it would appear that other less flexible but more specialised
techniques could be utilised to obtain analysis at a higher degree of granularity
of the data in question.
A summary of the overall performance of the techniques employed in this
paper are presented in Figure 1. The optimal LCA-ensemble result, while less
than the optimal NN-ensemble result, is obtained with somewhat less processing
effort and complexity, and further improvement may be possible.
Future work could look at extending the comparison of LCA with other
data mining algorithms to determine their applicability. Breast cancer represents
only one problem domain and applying these methods to other datasets would
be a logical extension. Our future research will include more experiments with
LatentGold R on other breast cancer datasets to determine how different numbers
of clusters produce different classification results for a more detailed analysis.
61
AIH 2012
8 Comparing Data Mining with Ensemble Classification
Fig. 1. Comparative Classification Accuracies.
References
1. Christoyianni, I., Koutras, A., Dermatas, E., Kokkinakis, G.: Computer Aided Diag-
nosis of Breast Cancer in Digitized Mammograms. Computerized Medical Imaging
and Graphics 26(5), 309-319 (2002)
2. DeSantis, C., Siegel, R., Bandi, P., Jemal, A.:Breast Cancer Statistics, 2011. CA: A
Cancer Journal for Clinicians 61(6), 408-418 (2011)
3. Fraley, C., Raftery, A.: Model-based Clustering, Discriminant Analysis, and Density
Estimation. Journal of the American Statistical Association 97(458), 611-631(2002)
4. Ghassem Pour, S., Maeder, A., Jorm, L.: Constructing a Synthetic Longitudinal
Health Dataset for Data Mining. DBKDA 2012, The Fourth International Confer-
ence on Advances in Databases, Knowledge, and Data Applications.86-90 (2012)
5. Ghassem Pour, S., Maeder, A., Jorm, L.: Validating Synthetic Health Datasets for
Longitudinal Clustering. The Australasian Workshop on Health Informatics and
Knowledge Management (HIKM 2013) 142, to appear (2013)
6. Gou, S., Yang, H., Jiao, L., Zhuang, X.: Algorithm of Partition Based Network
Boosting for Imbalanced Data Classification. The International Joint Conference
on Neural Networks (IJCNN).1-6. IEEE (2010)
7. Heath, M., Bowyer, K., Kopans, D., Moore, R., Kegelmeyer, P.: The Digital
Database for Screening Mammography. Proceedings of the 5th International Work-
shop on Digital Mammography.212-218 (2000)
8. Hofvind, S., Ponti, A., Patnick, J., Ascunce, N., Njor, S., Broeders, M., Giordano,
L., Frigerio, A., Tornberg, S.: False-positive Results in Mammographic Screening
for Breast Cancer in Europe: a literature review and survey of service screening
programmes. Journal of Medical Screening 19(1), 57-66 (2012)
9. Kim, J., Park, J., Song, K., Park, H.: Detection of Clustered Microcalssifications on
Mammograms Using Surrounding Region Dependence Method and Artificial Neural
Network. The Journal of VLSI Signal Processing 18(3),251-262 (1998)
62
AIH 2012
Comparing Data Mining with Ensemble Classification 9
10. Lanza, S., Flaherty, B., Collins, L.: Latent Class and Latent Transition Analysis.
Handbook of Psychology. 663-685 (2003)
11. Magidson, J., Vermunt, J.: Latent Class Models for Clustering: A Comparison with
k-means. Canadian Journal of Marketing Research 20(1), 36-43 (2002)
12. Malich, A., Schmidt, S., Fischer, D., Facius, M., Kaiser, W.: The Performance of
Computer-aided Detection when Analyzing Prior Mammograms of Newly Detected
Breast Cancers with Special Focus on the Time Interval from Initial Imaging to
Detection. European Journal of Radiology 69(3),574-578 (2009)
13. Mannila, H.: Data mining: Machine learning, Statistics, and Databases. Pro-
ceedings of Eighth International Conference on Scientific and Statistical Database
Systems.2-9 IEEE (1996)
14. McLeod, P., Verma, B.: Clustered Ensemble Neural Network for Breast Mass Clas-
siffcation in Digital Mammography. In: The International Joint Conference on Neu-
ral Networks (IJCNN). 1266-1271 (2012)
15. Mealing, N., Banks, E., Jorm, L., Steel, D., Clements, M., Rogers, K.: Investiga-
tion of Relative Risk Estimates from Studies of the Same Population with Contrast-
ing Response rates and Designs. BMC Medical Research Methodology 10(1), 10-26
(2010)
16. Mohanty, A., Senapati, M., Lenka, S.: A Novel Image Mining Technique for Clas-
sification of Mammograms Using Hybrid Feature Selection. Neural Computing &
Applications. 1-11 (2012)
17. Nishikawa, R., Kallergi, M., Orton, C., et al.: Computer-aided Detection, in its
present form, is not an Effective aid for Screening Mammography. Medical Physics
33(4), 811-814 (2006)
18. Nylund, K., Asparouhov, T., Muthen, B.: Deciding on the Number of Classes in
Latent Class Analysis and Growth Mixture Modeling: A Monte Carlo Simulation
Study. Structural Equation Modeling 14(4), 535-569 (2007)
19. Oh, S., Lee, M., Zhang, B.: Ensemble Learning with Active Example Selection for
Imbalanced Biomedical Data Classification. IEEE/ACM Transactions on Compu-
tational Biology and Bioinformatics 8(2), 316-325 (2011)
20. Sampat, M., Bovik, A., Markey, M.: Classification of Mammographic lesions into
BIRADS Shape Categories Using the Beamlet Transform. In: Proceedings of SPIE,
Medical Imaging: Image Processing. 16-25. SPIE(2005)
21. Surrendiran, B., Vadivel, A.: Feature Selection Using Stepwise ANOVA, Discrimi-
nant Analysis for Mammogram Mass Classification. International Journal of Recent
Trends in Engineering and Technology 3, 55-57 (2010)
22. Taylor, P., Potts, H.: Computer Aids and Human Second Reading as Interventions
in Screening Mammography: two systematic reviews to compare effects on cancer
detection and recall rate. European Journal of Cancer 44(6), 798-807 (2008)
23. Zhang, Y., Tomuro, N., Furst, J., Raicu, D.: Building an Ensemble System for
Diagnosing Masses in Mammograms. International Journal of Computer Assisted
Radiology and Surgery 7(2), 323-329 (2012)
63