Yazar "Abdulwahid, Ehab Saad" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim
    Öğe
    House price prediction using Artificial Neural Network (ANN) with adagrad optimizer
    (Altınbaş Üniversitesi / Lisansüstü Eğitim Enstitüsü, 2024) Abdulwahid, Ehab Saad; Ibrahim, Abdullahi Abdu
    The real estate market is a dynamic and complex ecosystem influenced by a myriad of factors, making accurate price predictions a formidable challenge. Understanding the intricate relationships between variables such as location, property characteristics, economic indicators, and market trends is essential for making informed investment decisions, in this thesis, a comprehensive exploration of machine learning and artificial neural networks (ANNs) has been undertaken, laying the groundwork for understanding how these powerful computational tools can be harnessed to solve complex problems across various domains. The study began by delving into the fundamentals of machine learning, categorizing it into its primary types, and discussing its applications in clustering, dimensionality reduction, and learning association rules. These sections highlighted the versatility and breadth of machine learning techniques in uncovering patterns and simplifying the complexities inherent in vast datasets. Further, a transition was made into a focused discussion on linear regression, including its simplest form and the more sophisticated gradient boosting method. This progression underscored the evolution of machine learning from basic predictive modelling to more advanced, iterative improvement techniques capable of handling nonlinear relationships with exceptional accuracy and efficiency.
  • [ X ]
    Öğe
    House price prediction using Artificial Neural Network with ADAGRAD optimizer
    (Springer Science and Business Media Deutschland GmbH, 2024) Abdulwahid, Ehab Saad; Ibrahim, Abdullahi Abdu; Abdulrazzaq, Mohammed Majid
    The real estate market is a dynamic and complex ecosystem influenced by myriad factors, making accurate price predictions a formidable challenge. Understanding the intricate relationships between variables such as location, property characteristics, economic indicators, and market trends is essential for making informed investment decisions. In this paper, we undertake a comprehensive exploration of machine learning and artificial neural networks (ANNs), establishing a framework to understand how these powerful computational tools can be harnessed to solve complex problems across various domains. The novel part of this study involves a comparative analysis of different optimization algorithms, including AdaGrad, Adam, SGD, and RMSprop, in the context of their application to ANN in real estate price prediction. AdaGrad’s unique approach to handling learning rates dynamically makes it particularly suited for data with varying scales, as demonstrated by our experimental results Comparative training and testing outcomes revealed that AdaGrad consistently outperformed other methods, showcasing lower mean squared error (MSE) and root mean squared error (RMSE), along with higher R2 values, indicating its superior predictive accuracy and efficiency. These findings underline the potential of AdaGrad to optimize real estate predictive models more effectively than traditional methods, marking a significant advancement in the application of machine learning techniques in real estate market analysis.