Volume:4 Issue:6 Jun ' 2018
||Enhance SVM Classification by Feature Weighting Function
Enhance SVM Classification by Feature Weighting Function
In recent years there are several changes and evolution has been done on data classification. The main role is assigning weights in feature classification learning algorithms. Assigning weights in classification learning called value weighting method. With the help of this method .It can assign different weight for each feature and also reduce the number of feature in dataset based on information gain and find the missing values in the dataset. Replace the missing value by using kullback-leibler method. This method is used to calculating the weights for missing values. The converted dataset will analysis the Support Vector Machine performance using number of datasets. The new paradigm results show that the Feature Weighting Function (FWF) method could improve the accuracy of Support Vector Machine significantly
||AN ASSESSMENT AND PREDICTION OF GSM NETWORK PERFORMANCE IN NIGERIA
AN ASSESSMENT AND PREDICTION OF GSM NETWORK PERFORMANCE IN NIGERIA
The paper carried out an assessment and prediction of Mobile Network Operators MNOs in Nigeria using the location marked by the longitudinal coordinates 6.6789oN and latitudinal coordinates 3.1666º E as a case study. Data was collected from the area marked by the longitude and latitude with the use of GSM phones. The collated data was analyzed using both sequential and statistical analysis in the excel environment and the results obtained from the analysis were forecasted over a period of time using linear regression analysis to show the most efficient mutual network. The results show poor connectivity between the network however, network D-H and B-F with an average Call Set-up Success Rate CSSR of 95.6% and 91.8% has the highest and lowest CSSR respectively. It is therefore envisaged that more work should be done to improve on the availability of the network in order to attain a sustainable development.