فهرست مقالات Gradient descent

• دسترسی آزاد مقاله
  • صفحه چکیده
  • متن کامل
  
  1 - Type-2 Fuzzy Logic Approach To Increase The Accuracy Of Software Development Effort Estimation
  Zahra Barati Mahdi Jafari Shahbazzadeh Vahid Khatibi Bardsiri
  predicting the effort of a successful project has been a major problem for software engineers the significance of which has led to extensive investigation in this area. One of the main objectives of software engineering society is the development of useful models to pre چکیده کامل

  predicting the effort of a successful project has been a major problem for software engineers the significance of which has led to extensive investigation in this area. One of the main objectives of software engineering society is the development of useful models to predict the costs of software product development. The absence of these activities before starting the project will lead to various problems. Researchers focus their attention on determining techniques with the highest effort prediction accuracy or on suggesting new combinatory techniques for providing better estimates. Despite providing various methods for the estimation of effort in software projects, compatibility and accuracy of the existing methods is not yet satisfactory. In this article, a new method has been presented in order to increase the accuracy of effort estimation. This model is based on the type-2 fuzzy logic in which the gradient descend algorithm and the neuro-fuzzy-genetic hybrid approach have been used in order to teach the type-2 fuzzy system. In order to evaluate the proposed algorithm, three databases have been used. The results of the proposed model have been compared with neuro-fuzzy and type-1 fuzzy system. This comparison reveals that the results of the proposed model have been more favorable than those of the other two models. پرونده مقاله
• دسترسی آزاد مقاله
  • صفحه چکیده
  • متن کامل
  
  2 - An Online Trajectory Tracking Control of a Double Flexible Joint Manipulator Robot by Considering the Parametric and Non-Parametric Uncertainty
  Alireza Pezhman Javad Rezapour Mohammadjavad Mahmoodabadi
  10.30495/admt.2021.1911965.1223
  Accurate trajectory tracking and control of the Double Flexible Joint Manipulator lead to design a controller with complex features. In this paper, we study two significant strategies based on improving the structure of the hybrid controller and training the controller چکیده کامل

  Accurate trajectory tracking and control of the Double Flexible Joint Manipulator lead to design a controller with complex features. In this paper, we study two significant strategies based on improving the structure of the hybrid controller and training the controller parameters for an online estimation of time-varying parametric uncertainities. For this purpose, combination of feedback linearization with an adaptive sliding mode control by considering update mechanism is utilized to stabilize the DFJM system. The update mechanism is obtained based on gradient descend method and chain rule of the derivation. Following, in order to eliminate the tedious trial-and-error process of determining the control coefficients, an evolutionary algorithm (NSGA-II) is used to extract the optimal parameters by minimizing the tracking error and control input. In the second step, an online estimation of the designed parameters were proposed based on three intelligent methods; weighting function, Adaptive Neural Network Function Fitting (ANNF), and adaptive Neuro-fuzzy inference system (ANFIS-PSO). The proposed controller reliability finally was examined in condition of the mass and the length of the robot arm was changed and sudden disturbances were imposed at the moment of equilibrium position, simultanously. The results of the tracking error and control input of the trained proposed controller demonstrated minimal energy consumption and shorter stability time in condition that the control parameters are constant and training are not considered. پرونده مقاله
• دسترسی آزاد مقاله
  • صفحه چکیده
  • متن کامل
  
  3 - A Hybrid Optimization Algorithm for Learning Deep Models
  Farnaz Hoseini Asadollah Shahbahrami Peyman Bayat
  Deep learning is one of the subsets of machine learning that is widely used in Artificial Intelligence (AI) field such as natural language processing and machine vision. The learning algorithms require optimization in multiple aspects. Generally, model-based inferences چکیده کامل

  Deep learning is one of the subsets of machine learning that is widely used in Artificial Intelligence (AI) field such as natural language processing and machine vision. The learning algorithms require optimization in multiple aspects. Generally, model-based inferences need to solve an optimized problem. In deep learning, the most important problem that can be solved by optimization is neural network training, but training a neural network can involve thousands of computers for months. In the present study, basic optimization algorithms in deep learning were evaluated. First, a performance criterion was defined based on a training dataset, which makes an objective function along with an adjustment phrase. In the optimization process, a performance criterion provides the least value for objective function. Finally, in the present study, in order to evaluate the performance of different optimization algorithms, recent algorithms for training neural networks were compared for the segmentation of brain images. The results showed that the proposed hybrid optimization algorithm performed better than the other tested methods because of its hierarchical and deeper extraction. پرونده مقاله
• دسترسی آزاد مقاله
  • صفحه چکیده
  • متن کامل
  
  4 - پیش بینی بازده بازار سرمایه با استفاده از الگوی یادگیری الگوریتم لورنبرگ مارکوات, گرادیان نزولی و الگوی آریما (ARIMA)
  مهدی اشعریون قمی زاده محمد محمودی
  پژوهش حاضر بر اساس ارزیابی الگوی یادگیری الگوریتم لورنبرگ مارکوات، گرادیان نزولی و الگوی آریما به مقایسه و توانایی پیش‌بینی کنندگی در بازار سرمایه می‌پردازد. بدین منظور داده‌های بازار در سال‌های 1394 تا 1397 مورد استفاده قرار گرفت و بیش از 75 درصد از این داده‌ها تا قبل چکیده کامل

  پژوهش حاضر بر اساس ارزیابی الگوی یادگیری الگوریتم لورنبرگ مارکوات، گرادیان نزولی و الگوی آریما به مقایسه و توانایی پیش‌بینی کنندگی در بازار سرمایه می‌پردازد. بدین منظور داده‌های بازار در سال‌های 1394 تا 1397 مورد استفاده قرار گرفت و بیش از 75 درصد از این داده‌ها تا قبل از سال 1397 به عنوان داده‌های آموزشی استفاده شد و داده‌های یک سال پایانی نیز به عنوان داده‌های آزمایشی مورد استفاده قرار گرفته شده است. نتایج تحقیق نشان داده‌اند، شبکه‌های عصبی مصنوعی ظرفیت بالایی برای پیش‌بینی قیمت دارند. مقایسه نتایج و عملکرد شبکه‌های عصبی و الگوی آریما (ARIMA) حاکی از آن است که شبکه عصبی قدرت پیش‌بینی بالاتری در مقایسه با الگوی خطی آریما (ARIMA) دارد، همچنین مقایسه عملکرد و دقت پیش‌بینی دو نوع شبکه عصبی با الگوریتم یادگیری لونبرگ مارکوارت و الگوریتم یادگیری گرادیان نزولی نشان داد که استفاده از الگوریتم یادگیری لونبرگ مارکورات توانسته است دقت پیش‌بینی شبکه عصبی را افزایش داده و خطای آن را کاهش دهد، بنابراین بر پایه پژوهش انجام شده می‌توان چنین نتیجه گرفت که الگوریتم یادگیری لونبرگ مارکوارت قدرت پیش‌بینی شبکه عصبی را بهبود می‌بخشد. پرونده مقاله