ANALYSIS OF USER MOBILITY PERFORMANCE ON SOFTWARE DEFINED WIRELESS NETWORK USING DIJKSTRA ALGORITHM

Over the last decade, wireless devices have developed rapidly until predictions will develop with high complexity and dynamic. So that new capabilities are needed for wireless problems in this problem. Software Defined Network (SDN) is generally a wire-based network, but to meet the needs of users in terms of its implementation, it has begun to introduce a Wireless-based SDN called Software Defined Wireless Network (SDWN) which provides good service quality and reach and higher tools, so as to be able to provide new capabilities to wireless in a high complexity and very dynamic. When SDN is implemented in a wireless network it will require a routing solution that chooses paths due to network complexity. In this paper, SDWN is tested by being applied to mesh topologies of 4,6 and 8 access points (AP) because this topology is very often used in wireless-based networks. To improve network performance, Dijkstra's algorithm is added with the user mobility scheme used is RandomDirection. The Dijkstra algorithm was chosen because it is very effective compared to other algorithms. The performance measured in this study is Quality of Service (QoS), which is a parameter that indicates the quality of data packets in a network. The measurement results obtained show that the QoS value in this study meets the parameters considered by the ITU-T G1010 with a delay value of 1.3 ms for data services and packet loss below 0.1%. When compared with the ITU-T standard, the delay and packet loss fall into the very good category.

INDOOR LOCALIZATION USING ACCELEROMETER AND GYROSCOPE SMARTPHONE BASED

Indoor localization is one of the more accurate technologies to be used to determine indoors or buildings. Pedestrian Dead Reckoning (PDR) is a method of determining the user's position by adding a method that occurs to a known initial position. The displacement that occurs is estimated with the help of an accelerometer sensor attached to the user as a step detector and to determine the direction towards the user using a gyroscope sensor. System testing is carried out in the Institut Teknologi Sumatera’s campus environment on the 2nd floor of Building C and D. The results from the detection of steps get an error rate of 1.13% using a threshold of 0.8.

SENTIMENT ANALYSIS OF THE COMMUNITY IN THE TWITTER TO THE 2020 ELECTION IN PANDEMIC COVID-19 BY METHOD NAIVE BAYES CLASSIFIER

The 2020 regional elections in the midst of the COVID-19 pandemic are starting to get crowded starting from the real world and in cyberspace, especially on Twitter social media. Twitter's existence has been widely used by various communities in recent years. Twitter is one of the media that represents the public response regarding public issu. Ahead of the general election (PEMILU), there are usually some parties who want to know the results of public sentiment or response to the issue, namely academics, intellectuals or even political opponents. Nevertheless, the implementation of local elections is very polemic in the community, therefore this study tries to analyze tweets that talk about issue public, namely the 2020 elections in the wake of the COVID-19 Pandemic. The analysis usually uses the classification of tweets containing public sentiment about the issue. The classification method used in this research is Naive Bayes Classifier (NBC) And Support Vector Machine (SVM). Naive Bayes Classifier is combined with features that can detect weighting using probability. The classification of tweets in this study was obtained based on a combination of two classes namely sentiment class and category class. The classification of sentiment consists of positive and negative. Test results on built-in applications show that accuracy with Naive Bayes delivers better results than Support Vector Machine. However, overall the use of the Naive Bayes method has a good performance to classify tweets with an accuracy rate of 92.2%

INDIVIDUAL IDENTIFICATION SYSTEM DESIGN THROUGH VOICE USING LINEAR PREDICTIVE CODING METHOD AND K-NEAREST NEIGHBOR

Humans have a variety of characteristics that are different from one another. Characteristics possessed by humans are genuine which can be used as a differentiator between one individual and another, one of which is sound. Voice recognition is called speech recognition. In this study, it was developed as an individual voice recognition system using a combination of the Linear Predictive Coding (LPC) method of feature extraction and K-Nearest Neighbor (K-NN) classification in the speech recognition process. Testing is done by testing changes in several parameters, namely the LPC order value, the number of frames, the K value, and different distance methods. The results of the parameter combination test showed a fairly good presentation of 73.56321839% with the combination parameter or LPC 8, the number of frames 480, the value of K 5, with the distance method used by Chebychev.

PREDICTION FOR COOPERATIVE CREDIT ELIGIBILITY USING DATA MINING CLASSIFICATION WITH C4.5 ALGORITHM

BMT Artha Mandiri is a cooperative that provides savings and loans services. In providing credit, BMT Artha Mandiri still uses the manual method, namely by looking at the ledger and history of each customer, to find out whether the applicant is worthy or not worthy of credit so that it is not effective and efficient. The purpose of this research is to make an application that can predict whether a prospective customer is eligible or not to be given credit. Predictions are made using the data mining classification method, namely the C4.5 algorithm based on the supporting data each customer has to classify which factors have the most influence on the level of credit payments in the cooperative. In a built application, the C4.5 algorithm produces a decision tree that is easy to interpret based on the existing variables. In the application, there are features that can be used to make decisions about customers who will apply for credit at the cooperative. The blackbox test results on the application show that the application has been able to run as expected, while the results of the algorithm test also show that the application has been able to implement the C4.5 algorithm correctly. In addition, the results of testing for accuracy show that the maximum average value of Accuracy is 79.19%.

THE CALCULATION SYSTEM OF NUMBER OF PEOPLE IN A ROOM BASED ON HUMAN DETECTION USING HAAR-CASCADE CLASSIFIER

Data of number of people in the room, calculations are usually carried out by assigning someone to oversee a room. In this final project, a system for calculating the number of people in the room is designed with image processing based on human detection that can be used in rooms, both for commercial applications and for security. This system uses Raspberry Pi device that already has an image processing method Haar-Cascade Classifier. Input data is in the form of video taken directly via webcam to be captured into a frame so that it can be used as a input the Haar-Cascade Classifier method and perform the counting process will be sent to the Antares platform. The system design has been tested with five scenarios. Scenario 1 the effect of the distance of the object, scenario 2 the effect of the pose of the object, scenario 3 the effect of the amount the object in the frame, scenario 4 affects the scale factor and scenario 5 measurement computation time. Scenarios 1 to 3 will do the best configuration for minimum neighbour. The system gets the best accuracy of 98,5% when the object distance 4 meters, the best accuracy of 96,6% when the object is facing forward and accuracy the best is 97,7% when the object in the frame is more than two objects with the best configuration use the minimum neighbour 5. Scenario 4 gets accuracy the best is 76,2% when using the scale factor 1.1. Scenario 5 gets the average computation time of the system is under one second, meaning the detection process done pretty fast.

WEBINAR STUDENT PRESENCE SYSTEM BASED ON REGIONAL CONVOLUTIONAL NEURAL NETWORK USING FACE RECOGNITION

World health organization announce Covid-19 as a pandemic so On March 15th 2020, the social distancing has been established with working, learning, and praying from home. Webinar is one of the solutions so those activities still can be done face to face and conference-based. With webinar, users can interact each other in an online meeting from home. Student presence is part of a webinar. The purpose of this research is to design an accurate student presence with a face recognition system using R-CNN method. The object of this research is a human face with sufficient light, medium, and the face must be facing the camera. This research proposed for a webinar student presence system is using face recognition with Regional Convolutional Neural Network (R-CNN). With object detection and several scenarios used in this method, the webinar student presence system using R-CNN will be more accurate than the methods that have ever been used before. This research has done four scenarios to obtain the best parameters like 45 of total layers, test data of the whole dataset percentage as 10%, RMSProp as model op- timizer, and 0.0001 learning rate. With those parameters, it have resulted the best system performance including 99.6% accuration, 1 × 10-4 loss, 100% precision, 99% recall, and 99.5% F1 Score.

IMPLEMENTATION TELEGRAM CHAT BOT ON STUDENT ORIENTATION PERIOD REGISTRATION SYSTEM FOR EFFICIENCY OF DATA MANAGEMENT

Orientation is a routine activity carried out by both public and private tertiary institutions. Universitas Muhammadiyah Surakarta (UMS) which is a private higher education institution also has several types of orientation, one of which is thePeriod of Ta'aruf the Muhammadiyah Student Association (MASTA IMM). IMM. IMM MASTA implementation that has occurred so far in the data processing process is still done manually, so there is often the same data and the time to manage the data is less effective and efficient. This study aims to design a MASTA IMM registration system using Telegram bot to simplify and streamline time in data management and class division. The method used in this research is themethod Waterfall modified, where the repair process is carried out only after the testing and evaluation stages. The development of this system will use the Telegram Bot API and theprogramming Pythonlanguage by utilizing library the provided. This system has several functions, including registration, adding and deleting data, dividing classes, and printing data intodocuments excel. This system has also been tested with good results. The test method used is the blackbox to find out the functionality of the system running properly. As well as testing the System Usability Scale (SUS) to evaluate the usefulness of the system with a final average result of 76.33, which means that the user agrees with the system that has been designed. This system is able to provide convenience during student registration and make it easier for admin in data management.

PEOPLE COUNTING FOR PUBLIC TRANSPORTATIONS USING YOU ONLY LOOK ONCE METHOD

People counting have been widely used in life, including public transportations such as train, airplane, and others. Service operators usually count the amount of passengers manually using a hand counter. Nowadays, in an era that most of human-things are digital, this method is certainly consuming enough time and energy. Therefore, this research is proposed so the service operator doesn't have to count manually with a hand counter, but using an image processing with You Only Look Once (YOLO) method. This project is expected that people counting is no longer done manually, but already based on computer vision. This Final Project uses YOLOv4 that is the latest method in detecting untill 80 classes of object. Then it will use transfer learning as well to change the number of classes to 1 class. This research was done by using Python programming language with various platforms. This research also used three training data scenarios and two testing data scenarios. Parameters measured are accuration, precision, recall, F1 score, Intersection of Union (IoU), and mean Average Precision (mAP). The best configurations used are learning rate 0.001, random value 0, and sub divisions 32. And the best accuration for this system is 69% with the datasets that has been trained before. The pre-trained weights have 72.68% of accuracy, 77% precision, and 62.88% average IoU. This research has resulted a proper performance for detecting and counting people on public transportations.

COMPARISON OF MACHINE LEARNING METHODS IN CLASSIFYING POVERTY IN INDONESIA IN 2018

Poverty is still one of the main problems in economic development besides inequality, unemployment, and economic growth. This study aims to model poverty directly using a discrete choice model, namely the machine learning classification method. The data used are imbalanced data where one of the categories is small enough so that the resample of both sampling method is used. In this study, several machine learning methods were applied, including the Decision Tree, Naïve Bayes, K-Nearest Neighbor (KNN), and Rotation Forest. The results show that the technique of using resample both samplings provides optimal results for the four machine learning methods. If viewed from the indicators of accuracy, specificity, sensitivity, AUC, and the highest Kappa coefficient produced, the best method is the KNN method. The KNN model has an accuracy value of 0.73 percent, sensitivity of 0.68 percent, specificity of 78 percent, and AUC of 0.73.