Browsing by Author "Musabe, Richard"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Hybrid model of Correlation based Filter Feature Selection and Machine Learning classifiers applied on Smart Meter Data set
    (IEEE, 2019) Sinayobye, Janvier Omar; Kaawaase Kyanda, Swaib; Kiwanuka, N. Fred; Musabe, Richard
    Feature selection is referred to the process of obtaining a subset from an original feature set according to certain feature selection criterion, which selects the relevant features of the dataset. It plays a role in compressing the data processing scale, where the redundant and irrelevant features are removed. Feature selection techniques show that more information is not always good in machine learning applications. Apply different algorithms for the data at hand and with baseline classification performance values we can select a final feature selection algorithm. In this paper, we propose a hybrid classification model, which has correlation based filter feature selection algorithm and Machine learning as classifiers. The objective of this study is to select relevant features and analyze the outperform machine learning algorithms in order to train our model, predict and compare their classification performance. In this method, features are ordered according to their Absolute correlation value with respect to the class attribute. Then top K Features are selected from ordered list of features to form a reduced dataset. This proposed classifier model is applied to our smart meter datasets. To measure the performance of these selected features; seven benchmark classifier are used; Random Forest (RF), Logistic Regression (LR), k-Nearest Neighbor (kNN), Naïve Bayes (NB), Decision Tree (DT), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM). This paper then analyzes the performance of all classifiers with feature selection in term of accuracy, sensitivity, F-Measure, Specificity, Precision, and MCC. From our experiment, we found that Random Forest classifier performed higher than other used classifiers.
  • Thumbnail Image
    Item
    PAPR reduction in LTE network using both peak windowing and clipping techniques
    (Journal of Electrical Systems and Information Technology, 2019) Musabe, Richard; Lionel, Mafrebo B.; Mugongo Ushindi, Victoire; Atupenda, Mugisha; Ntaganda, James; Bajpai, Gaurav
    Multicarrier technique orthogonal frequency division multiplexing (OFDM) modulation is a solution to provide high-speed and secured data transmission requirement in 4G technologies. Peak-to-average power ratio (PAPR) is one major drawback in OFDM system. Researches described several PAPR reduction techniques, notably peak windowing and clipping. The aim of this paper is to use these techniques to reduce PAPR. The research work describes clipping and windowing techniques such as quadratic amplitude modulation (QAM) and additive white Gaussian noise (AWGN) as channel condition. The simulation results show that in those techniques with clipping threshold level of 0.7, there is a reduction of PAPR of 8 dB, and the reduction of PAPR for the peak windowing when considering Kaiser window is about 11 dB.
  • Thumbnail Image
    Item
    Resources Allocation Optimization for Real Time Traffic in LTE Cellular Network
    (International Journal of Research in Electronics and Computer Engineering, 2016) Mukasine, Angelique; Ahishakiye, Faustin; Musabe, Richard; Bajpai, Gaurav
    Nowadays, a threefold real time traffic system is developing where voice, video and data traffic coexist in cellular networks. Since there are a limited number of resources, some of the traffic cannot be served in the real time. For example when all channels are occupied the remaining user equipments (UEs) requesting the service are blocked. Therefore a resource allocation optimization for real time traffic needs to be studied in this context. Significant studies have been conducted with a common objective of maximizing data rate and adaptive channel borrowing, however they don’t mitigate the problem of network coverage at the cell edge and spectrum is not shared efficiently among data and voice traffic. In this paper we propose three resource allocation methods for voice/data integrated in cellular network. In the First case, both voice and data traffic are assigned separate resources and the communication is in a dedicated mode. In the second case, we assign different resources but we allow the reuse of one of the traffic resources either voice or data provided that the interference is well managed. Thirdly we propose the method of borrowing the idle resources depending to the UE transmitting power of new call arrival. Our resource allocation schemes were developed based on blocking probability and weighted cell utilization mathematic model. MATLAB simulation results show that the performance of our proposed schemes in terms of UE blocking probability and cell weighted utility is better compared to cellular alone.