Puma: Permission Usage to detect Malware in Android - CISIS 2012

Editor's Notes

1. In statistics, logistic regression is a type of regression analysis used for predicting the outcome of a categorical (a variable that can take on a limited number of categories) criterion variable based on one or more predictor variables. The probabilities describing the possible outcome of a single trial are modelled, as a function of explanatory variables, using a logistic function.Logistic regression measures the relationship between a categorical dependent variable and usually a continuous independent variable (or several), by converting the dependent variable to probability scores.[1]
2. WehaveusedBayesiannetworksthatis a methodbasedontheBayes’ theoremWehavetheaforementionedfeatures.First, welearnthestructure of thenetwork, thatrepresenttherelationshipsbetweenthe variables. Thentheconditionalprobabilitytables are filled.There are different algoritmos tolearnthestructure of the net. WeusedK2, Hill Climber, TAN and a Naive net thatassumesthatall variables and independant
3. Decision Tree classifiers are a type of machine-learningclassifiers that are graphically represented as trees. Internalnodes represent conditions regarding the variables of a problem, whereas final nodes or leaves represent the ultimate decision of the algorithmDifferent training methods are typically used for learning thegraph structure of these models from a labelled dataset. Weuse Random Forest, an ensemble (i.e., combination of weakclassifiers) of different randomly-built decision trees [19], andJ48, the WEKA [20] implementation of the C4.5 algorithm
4. The K-Nearest Neighbour (KNN) [22] classifier is one ofthe simplest supervised machine learning models. This methodclassifies an unknown specimen based on the class of theinstances closest to it in the training space by measuringthe distance between the training instances and the unknowninstance.Even though several methods to choose the class of theunknown sample exist, the most common technique is tosimply classify the unknown instance as the most commonclass amongst the K-nearest neighbour
5. The K-Nearest Neighbour (KNN) [22] classifier is one ofthe simplest supervised machine learning models. This methodclassifies an unknown specimen based on the class of theinstances closest to it in the training space by measuringthe distance between the training instances and the unknowninstance.Even though several methods to choose the class of theunknown sample exist, the most common technique is tosimply classify the unknown instance as the most commonclass amongst the K-nearest neighbour
6. SVM algorithms divide the n-dimensional space representation of the data into two regions using a hyperplane. Thishyperplane always maximises the margin between those tworegions or classes. The margin is defined by the farthest distance between the examples of the two classes and computedbased on the distance between the closest instances of bothclasses, which are called supporting vectors [23].Instead of using linear hyperplanes, it is common to usethe so-called kernel functions. These kernel functions lead tonon-linear classification surfaces, such as polynomial, radialor sigmoid surfaces [24].
7. SVM algorithms divide the n-dimensional space representation of the data into two regions using a hyperplane. Thishyperplane always maximises the margin between those tworegions or classes. The margin is defined by the farthest distance between the examples of the two classes and computedbased on the distance between the closest instances of bothclasses, which are called supporting vectors [23].Instead of using linear hyperplanes, it is common to usethe so-called kernel functions. These kernel functions lead tonon-linear classification surfaces, such as polynomial, radialor sigmoid surfaces [24].