Person: GÜNAY, MELİKE
Loading...
Email Address
Birth Date
Research Projects
Organizational Units
Job Title
Arş.Gör.
Last Name
GÜNAY
First Name
MELİKE
Name
6 results
Search Results
Now showing 1 - 6 of 6
Publication Metadata only Digital Data Forgetting: A Machine Learning Approach(2018-10) Aşıroğlu, Batuhan; Yıldız, Eyyüp; Zencirli, Ahmet; Ensari, Tolga; Nalçakan, Yağız; GÜNAY, MELİKE; METE, BÜŞRA RÜMEYSA; 283334; 286270; 261185; 176400; 270484Digital transformation of the world goes very fast during last two decades. Today, data is power and very important. Firstly, magnetic tapes and then digital data storages have been used to collect all data. After this process, big data and its tool machine learning became very popular in both literature and industry. People use machine learning in order to obtain meaningful information from the big data. It brings valuable planning results. However, nowadays it is quite hard to collect and store all digital data to computers. This process is expensive and we will not have enough space to store data in the future. Therefore, we need and propose "Digital Data Forgetting" phrase with machine learning approach. With this digital / software solution, we will have more valuable data and will be able to erase the rest of the them. We called this operation "Big Cleaning". In this article, we use data set to get and extract more valuable data with principal component analysis (PCA), deep autoencoder and k-nearest neighbor machine learning methods in the experimental analysis section.Publication Metadata only Diagnosis of lung cancer using artificial immune system(International Scientific Meeting on Electrical-Electronics and Biomedical Engineering and Computer Science (EBBT), 2019) Orman, Zeynep; Ensari, Tolga; Oukid, Salina; Benblidia, Nadjia; GÜNAY, MELİKEIn this study, we implement the Artificial Immune System method to increase the number of data in the lung cancer dataset and obtain higher prediction rate for the diagnosis of lung cancer. Artificial Immune System is modified with the weights of features. Dataset dimension is decreased to raise the performance of this algorithm by using Pearson Correlation Coefficients. The system is also compared to other methods like k-Nearest Neighbor and Artificial Neural Networks that are commonly used in previous studies. As a result, the proposed weighted Artificial Immune System has the highest accuracy rate as 82% on normalized dataset and appears to be the second fastest method after k-Nearest Neighbor.Publication Metadata only New approach for predictive churn analysis in Telecom(2019) GÜNAY, MELİKE; 286270In this article, we propose a new approach for the churn analysis. Our target sector is Telecom industry, because most of the companies in the sector want to know which of the customers want to cancel the contract in the near future. Thus, they can propose new offers to the customers to convince them to continue using services from same company. For this purpose, churn analysis is getting more important. We analyze well-known machine learning methods that are logistic regression, Naïve Bayes, support vector machines, artificial neural networks and propose new prediction method. Our analysis consist of two parts which are success of predictions and speed measurements. Affect of the dimension reduction is also measured for the analysis. In addition, we test our new method with a second dataset. Artificial neural networks is the most successful as we expected but our new approach is better than artificial neural networks when we try it with data set 2. For both data sets, new method gives the better result than logistic regression and Naïve Bayes.Publication Metadata only Disease Prediction Using Weighted Artificial Immune System(2019-07) Orman, Zeynep; GÜNAY, MELİKE; 286270; 28475The Artificial Immune System (AIS) is a computational intelligence method inspired from the human immune system, which is applied to real-world problem solving related to classification, optimization and anomaly detection as an alternative approach to many data mining techniques. This paper presents a medical disease prediction system by using the AIS algorithm. The proposed system is implemented and tested on two different datasets which include breast cancer data and heart disease data with four different types of illness. Two other well-known data mining techniques that are Artificial Neural Networks (ANN) and K-Nearest Neighbor (KNN) are also tested on the same datasets to make a comparison in terms of their classification efficiency. By using AIS, accuracy obtained on breast cancer dataset is 98.08% and heart disease dataset is 70%. In addition to this, AIS algorithm gives the best classification results for both datasets. We also analyze the positive effect of preprocessing data before classification. Clearly, decreasing the number of different values that a class can be assigned for multivariate classes and assigning weights to each feature in heart disease dataset give prediction result with higher accuracy.Publication Metadata only Comparison of lung cancer detection algorithms(International Scientific Meeting on Electrical-Electronics and Biomedical Engineering and Computer Science (EBBT), 2019) GÜNAY, MELİKE; GÜNAYDIN, ÖZGE; ŞENGEL, ÖZNURLung cancer is a kind of difficult to diagnose and dangerous cancer. It commonly causes death both men and women so fast accurate analysis of nodules is more important for treatment. Various methods have been used for detecting cancer in early stages. In this paper, machine learning methods compared while detect lung cancer nodule. We applied Principal Component Analysis, K-Nearest Neighbors, Support Vector Machines, Naive Bayes, Decision Trees and Artificial Neural Networks machine learning methods to detect anomaly. We compared all methods both after preprocessing and without preprocessing. The experimental results show that Artificial Neural Networks gives the best result with 82,43% accuracy after image processing and Decision Tree gives the best result with 93,24% accuracy without image processing.Publication Metadata only Comparison of Face Recognition Algorithms(IEEE, 345 E 47Th St, New York, Ny 10017 USA, 2017) Ensari, Tolga; GÜNAY, MELİKE; 286270; 176400in this study, we analyze the algorithms that is used for face recognition and make performance comparison of two algorithms. The methods that is analyzed are k-nearest neighbors, Naive Bayes, eigenfaces, principle component analysis (PCA) and k-means are implemented on ORL face dataset. As a result of the analysis, k-nearest neighbors algorithm and eigenfaces algoritm are the most successful and Naive Bayes has the worst performance result. Performance of k-nearest neighborhood which is the most succesfull one is decreasing from %94 to %91.5 after the princible component analysis. In addition, the difference increases to %7 for Naive Bayes algorithm.