imkanlar
02
Dec 2020
BLU 696E Seminars

 
Name Surname Serdar Torun
Title Performance Analysis of Drone Cell Swarms Operating under JT-CoMP and unsupervised clustering
Abstract Drone base stations(DBS) are necessary in newly developing communication technologies. Thanks to their high line of sight possibility, the path loss of their transmitted signal is exremelylow. In highly crowded areas, multiple DBSs can serve to the users. In case of close flights, interference may occur for the cell edge users that stand between the coverage of two DBS. To prevent interference, joint transmission coordinated multipoint (JT-CoMP) technique is used.Deployment of DBSs in the field is decided with unsupervised clustering techniques k-means and Gaussian mixture models (GMM). The performance of deployment is compared in both techniques based on coverage probability.
Advisors Name and Surname Prof. Lütfiye Durak-Ata
Date 8.12.2020
Hour 13:30
Link https://itu-edu-tr.zoom.us/j/94474261290?pwd=bFlhQXZPNVRTSXJXeEJQd0J3RmFPdz09
   
Name Surname
 Kevser Şimşek
Title
 How to Position Machine Learning in Business?  A case : AIRLINE PASSENGER LOAD FACTOR PREDICTION
Abstract
 The aviation industry uses forecasting both to enable short term decisions, and to support longer term decisions in respect of future patterns in demand for air travel. The main aim of forecasting is to determine how patterns of demand will change over time, reflecting external factors such as growth in incomes, changes in prices and demographic changes. Forecasting is therefore a key tool for decision making, and is used in both business planning and policy decision making. With this study, it was aimed to forecast the passenger load factor (PLF) by using the information of two years reservation, group sales data, calendar information, weekly dates, trend difference between current year and previous year , past load factor information, load factor information of the same period of the previous year of Turkish Airlines which is a four star airline with a fleet of over 300 aircraft flying to over 290 destinations around the world. When each flight is thought to be its own characteristic, there is a need to find a solution for this work by a method that can reflect both the flight profile and the flight time dimension. Panel data regression method will be used for finding a solution of the problem. When the flight has not yet departed, a preliminary structure for both the economy and the business class will be obtained. In terms of revenue management, it is expected to optimize income, change capacities, efficiency of flight routes, forecasts for special days and certain flight days and months.
Advisors Name and Surname
  
Date
 15.12.2020
Hour
 14:30
Link
 https://itu-edu-tr.zoom.us/j/98383759614?pwd=UDNFNDNNeFR1cEtuS0dmRnl1UzZrUT09