Call Drops Research Articles

Design of Mobile Call Drop Reasons Prediction Model using Machine Learning

Call Drop (CD) is one of the most significant threats in today’s computing world since increasing usage of cellular networks. The call drops analysis and prevention methods is highly required for computer systems and avoiding call drops helps improve quality of service and customer satisfaction. This thesis is focused on descriptive prediction of the call drop reason using machine learning approaches and forecasting the call drop rate using ARIMA models. The raw data for this research were collected from Fault Management System, Addis Ababa for two years from 2020-21 mobile networks. The amount of data collected length is 59588 in number with six attribute were used for doing these research. The testing model used 11918, or 20%, of the dataset, while the training model used 47670, or 80%, of the dataset. To create descriptive classifications, six distinct modelling algorithms were utilised in this study. Those algorithms are: Logistic Regression (LR), Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Classification and Regression Trees (CART), Naive Bayes (NB), Support Vector Machine (SVM), the result shows that Decision Tree Classifier, shows good performance. With 57% accuracy, 53.2% precision, 49.2% recall, and 50% F1 score. The fitted number of call drop was calculated by optimum ARIMA (1,0,0) model with AIC value 2892.264 and RMSE value is 62.3819 and total model fit times is 7.280 seconds.

A Novel Chimp Optimized Linear Kernel Regression (COLKR) Model for Call Drop Prediction in Mobile Networks

Call failure can be caused by a variety of factors, including inadequate cellular infrastructure, undesirable system structuring, busy mobile phone towers, changing between towers, and many more. Outdated equipment and networks worsen call failure, and installing more towers to improve coverage might harm the regional ecosystems. In the existing studies, a variety of machine learning algorithms are implemented for call drop prediction in the mobile networks. But it facing problems in terms of high error rate, low prediction accuracy, system complexity, and more training time. Therefore, the proposed work intends to develop a new and sophisticated framework, named as, Chimp Optimized Linear Kernel Regression (COLKR) for predicting call drops in the mobile networks. For the analysis, the Call Detail Record (CDR) has been collected and used in this framework. By preprocessing the attributes, the normalized dataset is constructed using the median regression-based filtering technique. To extract the most significant features for training the classifier with minimum processing complexity, a sophisticated Chimp Optimization Algorithm (COA) is applied. Then, a new machine learning model known as the Linear Kernel Regression Model (LKRM) has been deployed to predict call drops with greater accuracy and less error. For the performance assessment of COLKR, several machine learning classifiers are compared with the proposed model using a variety of measures. By using the proposed COLKR mechanism, the call drop detection accuracy is improved to 99.4%, and the error rate is reduced to 0.098%, which determines the efficiency and superiority of the proposed system.

A Phenomenological Study of How Mobile Phone Users in India Cope with “Call Drops”

A Phenomenological Study of How Mobile Phone Users in India Cope with “Call Drops”

Prediction of Call Drops in GSM Network using Artificial Neural Network

Global System for Mobile communication is a digital mobile system that is widely used in the world. Over the years, the number of subscribers has tremendously increased, the quality of service (Call Drop Rate) became an issue to consider as many subscribers were not satisfied with the services rendered. In this paper, we present the Artificial Neural Network approach to predict call drop during an initiated call. GSM parameters data for the prediction were acquired using TEMS Investigations software. The measurements were carried out over a period of three months. Post analysis and training of the parameters was done using the Artificial Neural Network to have an output of “0” for no-drop calls and “1” for drop calls. The developed model has an accuracy of 87.5% prediction of drop call. The developed model is both useful to operators and end users for optimizing the network.

Optimizing Call Drops in Cellular Network using Artificial Intelligence based Handover Schema

Optimizing Call Drops in Cellular Network using Artificial Intelligence based Handover Schema

A Review of Menace of Call Drops in India and Possible Ways to Minimize It

In this age of wireless telecommunications, it is expected from the telecom service providers that the quality of services is the best; but unfortunately, it is not so and in India the telecom sector is struggling with a menace called Call Drops. Most of the operators in India do not meet the benchmark set by the TRAI for call drops. So, keeping in mind the Indian context of the problem, in this paper, the overall condition of call drops has been reviewed and possible ways to minimize the phenomena of call drops- Cell Splitting and Sectoring, Dynamic channel allocation, Hybrid channel allocation. Also, mathematical proof of relevance of these methods has been provided through Erlang B Formula and various other methods of proving used in other scientific pieces of work have been discussed such as Poisson Probability Function with a discrete variable. The data collected from various resources shows the gravity of the problem. Cell splitting and sectoring methods result in improvised soft handover mechanisms which in turn decrease the rate of call drops. This paper provides a brief but sufficed introduction to the term of call drops and also proposes some simple but efficient ways to cope this menace.

Reducing Call Drops Due to Cell Failure

Reducing Call Drops Due to Cell Failure

Investigation of Causes of Call Drops in GSM Network: A Case Study of Zain Nig. Ltd. in Benin City, Nigeria

This research is an investigation of the causes of call drops in a GSM network in Benin City, Nigeria. Call drop is inability of call to connect and go through the network successfully. For a drop call the subscriber does not enter into a queue for potential assignment. In this work, a drive test was conducted over a period of three months (March, April and May 2010) in cells around Benin City area. The drive test was conducted with the use of test mobile. A test mobile is a kit comprising mobile phone, drive testing proprietary software called TERMS on a laptop computer and GPS (global positioning system). A total of twenty seven cells out of forty cells around Benin City area were discovered to have severe call drops leading to very poor network performance. The problems that caused the call drops were determined and corrected leading to high improvement in call performance.

Prioritization Schemes in Queuing Handoff and New Calls to Reduce Call Drops in Cellular Systems

This paper proposes different queuing scenarios to avoid dropping of handoff and new calls in a cellular phone network, which is essential when the network has certain restrictions on the available frequencies. This limitation degrades the performance of the system and more sites are required to achieve the desired capacity and coverage. However, this leads to a higher percentage of call drops during handoff. This paper presents a queuing technique for both new and handoff calls to reduce the probability of call drop in such a system, leading to improvement in QoS. The proposed scenarios show better system performance. The blocking probability is reduced from 2% down to 0.04% for queuing handoff calls and from 2% to 1.14% when queuing new calls using the same technique. The four different presented approaches are: 1) only new calls are queued; 2) only handoff calls are queued; 3) by using all available channels; 4) by using only half. The queuing size also plays an important role for both new and handoff calls.