PPC did not help companies only on search engines, but also served as a billboard on various websites that are Display Network’s partners like Google Display Network, Yandex Ad Network etc. An amount that advertisers pay is calculated by predetermined factors by search engines, which are related to the bidding of competitors and quality of content published by the advertiser.Īfter the rise of online advertising, companies attempt to target specific audiences with PPC advertising solutions on search engines by using detailed targeting options. The PPC advertising, also known as cost-per-click advertising is a competition-based charging method, which is a click-based purchase model for advertisers and charges them for only clicks that reflect the visits to the links provided by the advertiser. This study focused on the role of automation and prediction about the future of a specific type of online advertising - pay-per-click advertising (PPC). Wilcoxon signed ranked test is conducted to demonstrate significant differences amid QDPSKNN and state-of-the-art methods. The results show improved classification performance with QDPSKNN in terms of precision, recall, f-measure, g-mean, reduction rate and execution time, compared to existing prototype selection methods in the classification of fraudulent publishers as well as on other benchmark imbalanced datasets. The performance is also compared with one baseline model (k-NN) and four other prototype selection methods such as NearMiss-1, NearMiss-2, NearMiss-3, and Condensed Nearest-Neighbor. The performance of QDPSKNN is evaluated on Fraud Detection in Mobile Advertising (FDMA) user-click dataset and fifteen other benchmark imbalanced datasets to test its generalizing behaviour. It reduces the size of the training dataset by selecting only the relevant prototypes in the form of nearest-neighbors.
The quad-division divides the data into four quartiles (groups) and performs controlled under-sampling for balancing class distribution. In this paper, we propose a Quad Division Prototype Selection-based k-Nearest Neighbor classifier (QDPSKNN) by introducing quad division method for handling uneven class distribution. Although Nearest-Neighbor techniques are simple to use and reduce the negative impact of the loss of potential information, they suffer from higher storage requirements and slower classification speed when applied on datasets with skewed class distributions. The Nearest-Neighbor techniques use Prototype Selection (PS) methods to select promising samples before classifying them for reducing the size of training data. The nearest-neighbor based classification techniques are popularly used to reduce the impact of class skewness on performance.
In online advertising, the user-clicks dataset based fraudulent publishers’ classification models exhibit poor performance due to high skewness in class distribution of the publishers.
0 kommentar(er)
