Intrusion Detection with Machine Learning and Feature Selection Methods

Kaynar, Oğuz; Arslan, Halil; Görmez, Yasin; Işık, Yunus Emre

doi:10.17671/gazibtd.368583

Cited by 21 publications

(10 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Kaynar et al. [19] obtained new datasets using attribute selection algorithms over a dataset developed for attack detection systems, and then using these new datasets k-nearest neighbor algorithm, support vector machines and over learning machine algorithms are applied and compared. As a result, they observed that feature selection methods increase the success rate of all three machine learning algorithms.…”

Section: Literature Reviewmentioning

confidence: 99%

Assessing countries’ performances against COVID-19 via WSIDEA and machine learning algorithms

Aydın

Yurdakul

2020

Applied Soft Computing

View full text Add to dashboard Cite

The COVID-19 pandemic, which first spread to the People of Republic of China and then to other countries in a short time, affected the whole world by infecting millions of people and have been increasing its impact day by day. Hundreds of researchers in many countries are in search of a solution to end up this pandemic. This study aims to contribute to the literature by performing detailed analyses via a new three-staged framework constructed based on data envelopment analysis and machine learning algorithms to assess the performances of 142 countries against the COVID-19 outbreak. Particularly, clustering analyses were made using k-means and hierarchic clustering methods. Subsequently, efficiency analysis of countries were performed by a novel model, the weighted stochastic imprecise data envelopment analysis. Finally, parameters were analyzed with decision tree and random forest algorithms. Results have been analyzed in detail, and the classification of countries are determined by providing the most influential parameters. The analysis showed that the optimum number of clusters for 142 countries is three. In addition, while 20 countries out of 142 countries were fully effective, 36% of them were found to be effective at a rate of 90%. Finally, it has been observed that the data such as GDP, smoking rates, and the rate of diabetes patients do not affect the effectiveness level of the countries.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Assessing countries’ performances against COVID-19 via WSIDEA and machine learning algorithms

Aydın

Yurdakul

2020

Applied Soft Computing

View full text Add to dashboard Cite

show abstract

“…formula where E is entropy, N is the total sample size, n is number of classes, and ns (i) is the sample for the i th class [54].…”

Section: Information Gainmentioning

confidence: 99%

The Attitudes of the Telecommunication Customers in the COVID-19 Outbreak: The Effect of the Feature Selection Approach in Churn Analysis

DONAT

USTA

2022

Sakarya University Journal of Science

View full text Add to dashboard Cite

Today's rising cutting-edge technology requirements and competitive environment in telecommunication industry has gained a remarkable importance due to the COVID-19 pandemics in terms of high need of information sharing and remote communication necessity. Telecommunication companies conduct significant analyses by highlighting that the customer data is the most valuable information. Besides, they obtain results emphasizing that acquiring new customers is costlier than retaining the existing ones. Therefore, the companies are willing to determine the important customer features in order to understand why they shift to the other telecommunication service providers. Hence, this study aims to conduct a churn analysis by feature selection approach with large volumes of telecommunication customer data in order to present what kind of customer behaviors and qualifications exist. Since there is a huge amount of data in this field, data mining is a vital requirement. The performance outputs were observed, and the features carrying these outputs to the highest value were identified. The data collection and analysis were carried out in mid-2019, and the same data collection and analysis were carried out again at the beginning of 2021, and these before and after results were compared. In addition, a comparison was made with the results obtained by the other churn analysis studies. This paper contributes to the practitioners by presenting the most important customer features in telecom customer churn, and a new approach in performance evaluation have been proposed specific to the telecommunication market with the industry experts’ guidance as a theoretical contribution.

show abstract

“…The SVM suggested by Cortes and Vapnik [14] uses the principle of structural risk minimization. SVM is a method of machine learning that divides data into two classes with the help of hyperplane [15]. SVM is one of the classifiers used for many different tasks, especially in recent years [16].…”

Section: Svmmentioning

confidence: 99%

“…In this method, it is tried to find a high plane where the distances of the nearest samples between the two classes are maximized. This method can often be used for linearly separating data; besides, it can also be used for nonlinearly separable data because it can make the data linearly separable with the help of the kernel functions [15]. The working principle of the classifier can be explained by an example; the data is classified by labeling the class to be detected as 1, and all other data in the other class as -1.…”

Section: Svmmentioning

confidence: 99%

Detecting of Warning Sounds in the Traffic using Linear Predictive Coding

Anacur

Saraçoğlu

2019

ijisae

View full text Add to dashboard Cite

Providing attendance of disabled persons into life and increasing it is an important social issue. In this context detecting sirens and sounds of those vehicles which have priority of way in traffic such as ambulance, fire-fighting vehicle and police car will enable to hearing disabled people to drive more comfortably. Recognising such warning sounds and detecting their direction have been aimed in this study. Sirens of ambulance, police car and fire-fighting vehicle in traffic have been classified as positive samples for application. Noises such as music and traffic noises have been classified as negative. Linear Estimator Coding has been made up by converting sound signals into digital data and qualities that reflect the sound in the best way has been determined by removing the qualities that do not reflect the sound in the data. By using principal component analysis method, meaningful and those qualities that represent the data in the best way have been classified by K Nearest Neighbor and Support Vector Machine Algorithm. After creating model about sound detection, model has been tested by performing new sound records. For the sound which has been detected to be warning sound, information about the sound direction has been given to user.

show abstract

Intrusion Detection with Machine Learning and Feature Selection Methods

Cited by 21 publications

References 23 publications

Assessing countries’ performances against COVID-19 via WSIDEA and machine learning algorithms

Assessing countries’ performances against COVID-19 via WSIDEA and machine learning algorithms

The Attitudes of the Telecommunication Customers in the COVID-19 Outbreak: The Effect of the Feature Selection Approach in Churn Analysis

Detecting of Warning Sounds in the Traffic using Linear Predictive Coding

Contact Info

Product

Resources

About