Loading...
Please wait, while we are loading the content...
Similar Documents
Detección de defectos en tiempo real en una línea de fabricación de tableros mediante métodos multivariantes de clasificación
| Content Provider | Semantic Scholar |
|---|---|
| Author | Collado, Alejandro Rodríguez |
| Copyright Year | 2019 |
| Abstract | The data analysis techniques have been gradually introduced in the world of factories, giving birth to the 4.0 Industry. Sonae Arauco has developed various devices that have allowed them to be multiple times awarded thanks to his technical innovation. They developed “Smart Eyes”, an imperfection-detecting system to be used in their board production line that works using filtering. The goal of this project is to create an imperfection-detecting system that uses data analysis techniques and Machine Learning. This End-of-Degree Project is of a double degree (INDat – double degree in Computer Engineering and Statistics), so it consists of two memories, one per degree. This is the memory corresponding to Statistics degree and it consists of several parts. The first part will describe the general state of the problem: data obtained from the company, its characteristics and possible useful features to extract for the posterior data analysis. The second part will describe different classifiers as well as its results. The classifiers appearing in this part will be related to discriminant analysis, decision trees and its extensions. In the Computer Engineering End-of-Degree Project memory, you can find classifiers related to neural networks and support vector machines. The third part will analize the obtained results of each classifier. This will allow us to know what has been the method that has obtained the best results. The performance of each classifier will be written on the memory that has previously explained the theory behind the classifier. The epilogue will serve to explain the observed conclusions related to the achivement of proposed goals as well as break down the possible future work. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://uvadoc.uva.es/bitstream/handle/10324/35997/TFG-G3536.pdf?isAllowed=y&sequence=1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
