- An Analysis of Android Malware and IoT Attack Detection with Machine Learning
Keywords
Internet of Things, Security, Malware Attacks, Android Malware dataset, Machine Learning
An identification process of IoT malware attacks is carried out in our study by employing various ML methods such as Naive Bayes (NB), K-Nearest Neighbour (KNN), Decision Tree (DT), and Random Forest (RF). We constructed a ML framework by utilizing a set of Android malware data samples and various good applications. Results show that, Decision Tree algorithm achieved greater outcomes than others.
- Machine Learning Approaches for Analysing Static features in Android Malware Detection
Keywords
Android, Malware, CICInvesAndMal2019, Random Forest (RF), K-NN, Naive Bayes, Trojan, Ransomware, Adware
In the process of Android malware detection, we utilized Principal Component Analysis (PCA) to select the relevant features. We trained and examined the dataset by employing several ML techniques including Naive bias, Decision tree (DT), Random Forest (RF), and k-NN. From the analysis, we conclude that, Random forest method is considered as an optimal classifier in both binary and category classification.
- A Novel Machine Learning Approach for Android Malware Detection Based on the Co-Existence of Features
Keywords
Co-existence, FP-growth
A machine learning framework for the purpose of android malware identification is recommended in our study that is related to co-existence of static features. We extracted the important co-existed features by utilizing association rule mining based method named FP-growth technique. We examined our suggested framework by employing traditional ML approaches. At last, RF method and co-existence of features provides efficient outcomes.
- Malware Detection in Android Based Devices by a Hybrid Approach Using Machine Learning Techniques
Keyword
Ensemble Learning
ML based android malware detection models are reviewed in our research. Here we discussed about framework of android apps, safety approaches, android model patterns and classification of malware. We described various procedures that exist in many research studies, they are data preprocessing, feature selection, ML frameworks and techniques and identification accuracy. At last, various future hazards in android malware identification are evaluated.
- A Study on Android Malware Detection Using Machine Learning Algorithms
Keywords
Android malware, Malware detection, Permissions, Activities, Smartphone protection
We constructed an identification model based on ML approaches by integrating various application characteristics denoted as permission and activities that are acquired at the time of app installation. We utilized Androguard tool to extract permissions and activities of every apps. After that, we categorize the apps as benign or malicious by using these extracted features. As a result, RF provides better performance and KNN provides least performance among others.
- Android Malware Detection Using Machine Learning Classifiers
Keywords
Android malware detection, Feature selection
Our article proposed a ML based model for android malware identification. We performed random sampling to retrieve a balanced dataset from CICAndMal2017 dataset. Then we obtained the most relevant features by employing feature engineering process. By utilizing the chosen features, we trained the ML techniques. As a consequence, Random Forest technique achieved greater end results.
- Detection of Android Malwares on IOT Platform Using PCA and Machine Learning
Keywords
Dynamic features, Static features, PCA, Deep learning
For malware detection analysis, we utilized real devices such as android smart phones, watches and tablets in our study rather than using genymotion. We extracted the features from normal and harmful applications. To minimize the size of dataset, we employed PCA technique to extract only the relevant features. Then we input the extracted features into DL framework with various hidden layers. However, our suggested framework offers best end results.
- A Comparative Analysis of Machine Learning Algorithms for Android Malware Detection
Keyword
Information Security
A comparative analysis for various ML approaches is carried out in our work and their performances related to android malware identification are evaluated. We normalize the features by using Synthetic Minority Oversampling Technique (SMOTE). We also utilized PCA approach in our work. We constructed a LGBM framework to detect and categorize the android malware into different categories.
- An Ensemble Approach Based on Fuzzy Logic Using Machine Learning Classifiers for Android Malware Detection
Keywords
Mobile security, fuzzy logic
We suggested a fuzzy logic-based dynamic ensemble (FL-BDE) framework to identify the android malware. Our framework comprises of ML processing power and Mamdani-type fuzzy inference system decision making power. We employed various ML methods like LR, BP), BDT, NN, DF and SVM. We performed the experimental analysis for our suggested model and ML techniques and it is considered as an efficient malware detection approach.
- DynaMalDroid: Dynamic Analysis-Based Detection Framework for Android Malware Using Machine Learning Techniques
Keywords
Vulnerabilities, Dynamic analysis, System calls
A dynamic analysis related model named DynaMalDroid is recommended in our research for the detection of harmful applications exists in android framework. By using dynamic analysis, various apps are extracted. We selected the relevant features by using feature engineering method. Several ML methods like RF, DT, LR, SVM, NB, KNN, and AdaBoost are employed to detect the malicious applications. In that, AdaBoost and SVM outperform others.