کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
5125077 | 1488271 | 2017 | 7 صفحه PDF | دانلود رایگان |
کلمات کلیدی
1.مقدمه
2. پیشینه ی نظری
1.2. سیستم های اطلاعات جغرافیایی
2.2. خوشه بندی فضایی
3.2. شبکه های عصبی مصنوعی
شکل 1. یک شبکه عصبی پیش نگر نوعی. I1,I2, …,In جزء مقادیر ورودی و Q1,Q2,…,Qn جزء مقادیر خروجی هستند.
4.2. الکوریتم گرادیان مزدوج مقیاس پذیر
3. روش تحقیق
شکل 2. مروری بر روش تحقیق اعمال شده
4. نتایج
1.4. جمع آوری و تهیه داده
2.4. تحلیل و خوشه بندی فضایی
3.4. مدل های پیش بینی شبکه عصبی مصنوعی
شکل 3. عملکرد آموزشی شبکه عصبی مدل بهینه براساس خطای میانگین مربعات (MSE)
4.4 پیش بینی های مربوط به مدل شبکه عصبی مطلوب
5. بحث و نتیجه گیری
Public administration has adopted information and communication technology in order to construct new intelligent systems and design new risk prevention strategies in transportation management. The ultimate goal is to improve the quality of the transportation services and also to ensure public transportation safety. In this research, a combination of spatial clustering methods and artificial neural network models was used in order to predict the high crime risk transportation areas. Geographic information systems were used to perform spatial analysis so as to identify the regions with a high concentration of crime incidents. Artificial intelligence was used in this study in order to build artificial neural network predictive models. The neural network predictive models were evaluated by using the Mean Squared Error (MSE) in order to find the optimal forecasting model. The optimal forecasting model was used in order to predict the high crime risk transportation areas. The scaled conjugate gradient algorithm was utilized as the training algorithm for the construction of the feedforward neural network models, since it is considered as one of the fastest learning algorithms compared to several other algorithms such as backpropagation learning algorithms.
Journal: Transportation Research Procedia - Volume 24, 2017, Pages 467-473