Scribd
Upload
23 views14 pages

Results Thesis

Uploaded by

vijaygate2013
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
23 views14 pages

Results Thesis

Uploaded by

vijaygate2013
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 14

to prove the effectiveness of the model modified ABC algorithm applied on both OASIS

longitudinal and ADNI Biomarkers datasets. it is high dimensional dataset.

stat related oasis longitude

Data set description


Cycle No of Accuracy of
features proposed
selected Ensemble
classification
model
1 5 92.00000167
2 4 80
3 8 92.00000167
4 8 92.00000167
5 6 92.00000167
6 6 78.66666667
7 3 64
8 4 92.00000167
9 5 70.66666667
10 3 61.33333333
11 6 70.66666484
12 6 81.33333325
13 7 92.00000167
14 5 92.00000167
15 3 57.33333333
16 4 93.33333333
100
90
80
70
60
Cycle
50 No of features selected
40 Accuracy of proposed En-
semble classification model
30
20
10
0
1 2 3 4 5 6 7 8 9 10111213141516

AD biomarker
Results over Biomarker data set

The model designed and evaluated using a variety of performance matrix compare with
traditional models our proposed model produce better results with high accuracy and less time.
The model performance are evaluated in terms of accuracy, specificity, recall, precsion, f1
score. In our model performed in an iterative process on each iteration Modified Artificial Bee
Colony algorithm produce set of optimised features and then build the multi-class based random
forest. The model is tested over the test data to achieve better accuracy. The model build as an
iterative approach using convergency to achieve better accuracy.
Table 5.1 Cycles and Accuracy of a Proposed method

Table 5.1 Depicts the No of Cycles and Accuracy of Proposed Iterative Approach of Random
Forest Classifier with an Artificial Bee Colony Algorithm

Cycle No of features selected Random forest Accuracy


0 36 0.9771750805585392
1 49 0.978249194414608

2 44 0.9511278195488722
3 30 0.9723415682062299
4 47 0.952207933404941
5 43 0.9774436090225563
6 41 0.943875570204082
7 42 0.9575725026852846
8 51 0.9742212674543801
9 39 0.9527389903329753
10 45 0.970130397421267
11 49 0.9742212674543801
12 43 0.9669709989258861
13 40 0.9487110633727115
14 56 0.9691192266380236
15 42 0.9575725026852846
16 44 0.924818030098783
17 34 0.9401181525241675
18 49 0.9618689581098896
19 40 0.9747583243823545
20 49 0.9616004296455424
21 43 0.9750268528464017
22 55 0.9683136412459729
23 44 0.9578410311493017
24 51 0.9701933404940923
25 49 0.9830827067669173
120

100

80

Cycle
60 No of features selected
accuracy of proposed En-
40 semble classification model

20

0
1 4 7 10 13 16 19 22 25

Based on the iterative cycles report, observed that when proposed model produce less
than 42 features within 96 features are generate with low performance in terms of
accuracy.
In each iteration cycle, it selects subset of features and model generate the
accuracy as 98.08 in 25 th cycle with 49 subsets of features. Based on the iterative
process by observing various cycles which are produced more than 45 to 49 features
are produced better performance in terms of accuracy.
Table 5.2Analysis
Table 5.2 Depicts the Existing Technique and Proposed Approach Analysis.

Author Technique Classification


Accuracy
Xulong Wang[1] DT 79%
Akash Shah[2] Machine learning 81%
algorithms
Morshedul Bari Machine learning 81.3%
Antor[3] algorithms
Mumine Kaya Keles[4] BABC 83.9%
Randa Ben Ammar[5] Svm-RFE 85%
Malavika.G[6] Machine learning 86.8%
algorithms
Ji Hwan Park[7] SABC 89.13%
D. Chitradevi[8] ABC 90%
Guilherme Tavares[9] ECG 90%
Minseok Song[10] Machine learning 90.2%
algorithms
M. Tanveer[11] RF-RSVM 90.53%
K.G.Achilleos[12] Data 91.4%
Augmentation
Afreen Khan[13] CML 93.90%
Moolchand Sharma[14] ABC 95.71%
Yusera Farooq XGB + DT + 95.75%
Khan[15] SVM using Grid
Search
Our Proposed Approach MABC 98.30%

In the fig 5.3 depicts with the use of training data, a proposed iterative approach of random
forest classification is trained for cycles in every approach. But in the 25th cycle the proposed
approach selects the 49 features with an accuracy of 0.9830827067669173.
model2

data set description


model3
dataset description

selected features

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy