Dynamic Regressor/Ensemble Selection for a Multi-Frequency and Multi-Environment Path Loss Prediction

نویسندگان

چکیده

Wireless network parameters such as transmitting power, antenna height, and cell radius are determined based on predicted path loss. The prediction is carried out using empirical or deterministic models. Deterministic models provide accurate predictions but slow due to their computational complexity, they require detailed environmental descriptions. While less accurate, Machine Learning (ML) fast with accuracies comparable that of Most Empirical versatile valid for various values frequencies, heights, sometimes environments, whereas most ML not. Therefore, developing a model will surpass accuracy entails collecting data from scenarios different environments the develop model. Combining datasets sizes could lead lopsidedness in particular scenario low imbalance. This because varies at certain regions dataset variations more intense when generated fusion sizes. A Dynamic Regressor/Ensemble selection technique proposed address this problem. In method, regressor/ensemble selected predict sample point sample’s proximity cluster assigned regressor/ensemble. K Means Clustering was used form clusters regressors considered Nearest Neighbor (KNN), Extreme Trees (ET), Random Forest (RF), Gradient Boosting (GB), (XGBoost). ensembles any combinations two, three four regressors. points belonging each were validation set regressor made lowest absolute error per individual point. Implementation resulted improvements described by few training data. Improvements also observed other works compared reported works. study shows features extracted satellite images describe environment appropriate than categorical clutter height value.

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

tight frame approximation for multi-frames and super-frames

در این پایان نامه یک مولد برای چند قاب یا ابر قاب تولید شده تحت عمل نمایش یکانی تصویر برای گروه های شمارش پذیر گسسته بررسی خواهد شد. مثال هایی از این قاب ها چند قاب های گابور، ابرقاب های گابور و قاب هایی برای زیرفضاهای انتقال پایاست. نشان می دهیم که مولد چند قاب تنک نرمال شده (ابرقاب) یکتا وجود دارد به طوری که مینیمم فاصله را از ان دارد. همچنین مسایل مشابه برای قاب های دوگان مطرح شده و برخی ...

15 صفحه اول

Multi-Stage Fuzzy Load Frequency Control Based on Multi-objective Harmony Search Algorithm in Deregulated Environment

A new Multi-Stage Fuzzy (MSF) controller based on Multi-objective Harmony Search Algorithm (MOHSA) is proposed in this paper to solve the Load Frequency Control (LFC) problem of power systems in deregulated environment. LFC problem are caused by load perturbations, which continuously disturb the normal operation of power system. The objectives of LFC are to mini small size the transient deviati...

متن کامل

Multi Agent Path Planning Approach to Dynamic Free Flight Environment

For years the air traffic routes have been fixed by the air traffic controllers on the basis of pre-acquired knowledge. Free flight air traffic management is a solution to move the decision making process for route choices from air traffic controllers to the cockpit. The basic idea is optimize the flight routes in the consideration of multiple objectives .i.e. the optimal path is decided based ...

متن کامل

A Fuzzy Approach For Multi-Objective Supplier Selection

Assessment and selection of suppliers are two most important tasks in the purchasing part in supply chain management. Supplier selection can be considered to be a single or multi-objective problem. From another point of view, it can be a single or multi-sourcing problem. In this paper, an integrated AHP and Fuzzy TOPSIS model is proposed to solve the supplier selection problem. This model makes...

متن کامل

A fuzzy random multi-objective approach for portfolio selection

In this paper, the portfolio selection problem is considered, where fuzziness and randomness appear simultaneously in optimization process. Since return and dividend play an important role in such problems, a new model is developed in a mixed environment by incorporating fuzzy random variable as multi-objective nonlinear model. Then a novel interactive approach is proposed to determine the pref...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Information

سال: 2022

ISSN: ['2078-2489']

DOI: https://doi.org/10.3390/info13110519