Support Vector Machine Accuracy Improvement with Classification

Mohan, Lalit; Pant, Janmejay; Suyal, Priyanka; Kumar, Arvind

doi:10.1109/cicn49253.2020.9242572

Cited by 70 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The Kernel Trick is a technique that is used by SVM for the transformation of low-dimension data into highdimension data. The SVM is developed with various kernel tricks, but the most common kernel tricks are linear, polynomial, radial basis function (RBF), and Sigmoid (Mohan et al, 2020), (Alam et al, 2016)…”

Section: Traditional ML Methodsmentioning

confidence: 99%

Efficient and Low-Cost Skin Cancer Detection System Implementation with a Comparative Study Between Traditional and CNN-Based Models

Mampitiya,

Rathnayake,

De Silva

2022

JCCE

View full text Add to dashboard Cite

Medical image classification is an essential task in the field of combining medical applications with Artificial Intelligence. This study is carried out to introduce an accurate, precise method for skin cancer recognition. This research investigates the performance of classifying skin cancer dataset HAM10000 using ResNet50, MobileNet, and the traditional Support Vector Machines (SVM) model. The dataset combines seven cancer types: Actinic Keratosis, Basal Cell Carcinoma, Benign Keratosis, Dermatofibroma, Melanoma, Melanocytic Nevus, and Vascular Lesion. The SVM classifier is designed to employ a Histogram of Oriented Gradient (HOG) features with Principle Component Analysis (PCA). Moreover, the Synthetic Minority Oversampling Technique is used to balance the dataset. Additionally, six conventional machine learning (ML) methods are used to compare the results with the calculation of precision, recall, F1 Score, and accuracy. The results confirm that the SVM method outperforms the other algorithms with an accuracy of 99.15%. The novelty contribution of this research activity is mainly based on the development of a high accuracy, low computational complex machine method for skin cancer types recognition in the domain of medical image classification.

show abstract

Section: Traditional ML Methodsmentioning

confidence: 99%

Efficient and Low-Cost Skin Cancer Detection System Implementation with a Comparative Study Between Traditional and CNN-Based Models

Mampitiya,

Rathnayake,

De Silva

2022

JCCE

View full text Add to dashboard Cite

show abstract

“…An SVM identifies the optimal hyperplane to separate two signals in the dataset. Traditional SVM has been used in several problems of engineering, such as network security, data classification, and bioinformatics [31].…”

Section: Support Vector Machinementioning

confidence: 99%

“…A linear kernel is used when the function tends to be linearly separable and uses small samples and the training process is short. Another kernel, such as a polynomial, can be used when the function is not linearly separable and uses variables such as the polynomial degree that can affect the computational time and results [31].…”

Section: Support Vector Machinementioning

confidence: 99%

Machine Learning Techniques Based on Primary User Emulation Detection in Mobile Cognitive Radio Networks

Muñoz

Martínez

Hernández

2022

Sensors

View full text Add to dashboard Cite

Mobile cognitive radio networks (MCRNs) have arisen as an alternative mobile communication because of the spectrum scarcity in actual mobile technologies such as 4G and 5G networks. MCRN uses the spectral holes of a primary user (PU) to transmit its signals. It is essential to detect the use of a radio spectrum frequency, which is where the spectrum sensing is used to detect the PU presence and avoid interferences. In this part of cognitive radio, a third user can affect the network by making an attack called primary user emulation (PUE), which can mimic the PU signal and obtain access to the frequency. In this paper, we applied machine learning techniques to the classification process. A support vector machine (SVM), random forest, and K-nearest neighbors (KNN) were used to detect the PUE in simulation and emulation experiments implemented on a software-defined radio (SDR) testbed, showing that the SVM technique detected the PUE and increased the probability of detection by 8% above the energy detector in low values of signal-to-noise ratio (SNR), being 5% above the KNN and random forest techniques in the experiments.

show abstract

“…Untuk model Decision Tree saja saat ini sudah ada algoritma hasil pengembangannya yaitu Random Forest yang dikembangkan pertama kali oleh Breiman pada penelitian [15]. Selain pendekatan model pohon keputusan seperti dua metode yang telah disebutkan, masih ada model lain yang menggunakan pendekatan mathematical modelling seperti support vector machine [16] [17]. Algoritma klasifikasi lain adalah Naïve Bayess yang menggunakan konsep probabilitas dalam proses klasifikasinya [18].…”

Section: Klasifikasiunclassified

Seleksi Fitur dan Penanganan Imbalanced Data menggunakan RFECV dan ADASYN

Pratama¹,

Chandra²,

Presetyaningrum³

2022

eksplora

View full text Add to dashboard Cite

Proses data mining bekerja terhadap data yang tersedia. Jika dataset tidak tersedia sepenuhnya, hasil pengolahan data mining menjadi tidak optimal. Terdapat beberapa kondisi data yang perlu penanganan terlebih dahulu sebelum memasuki tahap data mining. Salah satunya ialah imbalanced class yang merupakan kondisi di mana distribusi data pada setiap kelas tidak proporsional. Sebagai salah satu cara untuk efisiensi proses klasifikasi, seleksi fitur dapat memenuhi kebutuhan tersebut karena hasil dari seleksi fitur adalah sebuah dataset dengan jumlah atribut yang lebih sedikit dari sebelumnya. Untuk menyelesaikan permasalahan imbalanced class, ADASYN digunakan dalam penelitian ini sebagai metode untuk menyeimbangkan proporsi kelas pada dataset. Sedangkan RFECV digunakan sebagai metode fitur seleksi yang dapat meningkatkan efisiensi pada proses klasifikasi. Setelah dilakukan evaluasi dari hasil klasifikasi pada dataset yang menggunakan seleksi fitur, didapatkan hasil klasifikasi yang lebih baik dibandingkan dengan hasil klasifikasi pada dataset tanpa seleksi fitur. Hal tersebut dibuktikan dengan perbandingan antara hasil terbaik dari akurasi klasifikasi dataset tanpa seleksi fitur. Hasil dari metode CART sebesar 85.1% yang merupakan hasil dari pengolahan data tanpa menggunakan metode fitur seleksi. sedangkan metode Bagging k-NN yang menghasilkan akurasi sebesar 88% yang di aplikasikan pada dataset dengan seleksi fitur. Sehingga dapat disimpulkan bahwa seleksi fitur dapat meningkatkan akurasi pada klasifikasi.

show abstract

Support Vector Machine Accuracy Improvement with Classification

Cited by 70 publications

References 14 publications

Efficient and Low-Cost Skin Cancer Detection System Implementation with a Comparative Study Between Traditional and CNN-Based Models

Efficient and Low-Cost Skin Cancer Detection System Implementation with a Comparative Study Between Traditional and CNN-Based Models

Machine Learning Techniques Based on Primary User Emulation Detection in Mobile Cognitive Radio Networks

Seleksi Fitur dan Penanganan Imbalanced Data menggunakan RFECV dan ADASYN

Contact Info

Product

Resources

About