| (2 intermediate revisions by 2 users not shown) | |||
| Line 3: | Line 3: | ||
</font size> | </font size> | ||
| − | A [https://www.projectrhea.org/learning/slectures.php slecture] by Chiho Choi | + | A [https://www.projectrhea.org/learning/slectures.php slecture] by [[ECE]] student Chiho Choi |
| − | + | ||
| + | Partly based on the [[2014_Spring_ECE_662_Boutin|ECE662 Spring 2014 lecture]] material of [[user:mboutin|Prof. Mireille Boutin]]. | ||
</center> | </center> | ||
---- | ---- | ||
| Line 16: | Line 17: | ||
=Questions and Comments= | =Questions and Comments= | ||
| − | * | + | * (Reviewed by Sang Ho Yoon) The author introduces the density estimation using Parzen window as well as classification result based on the Parzen window estimation. It goes over the basic mathematical approach to the Parzen window method. The Parzen window on various dimension is well demonstrated with figures. The author explains in detail about effect of parameters including window width and sample sizes. All results are clearly demonstrated in figures. It is mentioned that the window width selection is very import when the sample size is small. If sample sizes are large enough, the window width will have low impact on the density estimation. The report also covers the convergence of the mean and variance as sample size grows to the infinity. In last part, it explains about relationship between sample size and classification error based on Parzen window estimation. It concludes with pros and cons of using Parzen Window.<br> |
| − | + | * (Reviewed by Sang Ho Yoon) Overall, the report is well written and explained in a sense that it enhances an understanding of mathematical proof by matching experimental results. With the aid of graphical visualization, it is easy to follow through the lecture. Author's opinion on Parzen window in dicussion section is quite useful since it is based on the author's experience on various experiments on this method. <br> | |
| + | * (Reviewed by Sang Ho Yoon) I recommend the author to perform analysis on relationship between window width and sample sizes. The report could get benefits from testing various kernel functions. <br> | ||
---- | ---- | ||
Latest revision as of 08:32, 1 May 2014
Questions and comments
A slecture by ECE student Chiho Choi
Partly based on the ECE662 Spring 2014 lecture material of Prof. Mireille Boutin.
If you have any questions, comments, etc. please post them on this page.
Go to Parzen window method and classification.
Questions and Comments
- (Reviewed by Sang Ho Yoon) The author introduces the density estimation using Parzen window as well as classification result based on the Parzen window estimation. It goes over the basic mathematical approach to the Parzen window method. The Parzen window on various dimension is well demonstrated with figures. The author explains in detail about effect of parameters including window width and sample sizes. All results are clearly demonstrated in figures. It is mentioned that the window width selection is very import when the sample size is small. If sample sizes are large enough, the window width will have low impact on the density estimation. The report also covers the convergence of the mean and variance as sample size grows to the infinity. In last part, it explains about relationship between sample size and classification error based on Parzen window estimation. It concludes with pros and cons of using Parzen Window.
- (Reviewed by Sang Ho Yoon) Overall, the report is well written and explained in a sense that it enhances an understanding of mathematical proof by matching experimental results. With the aid of graphical visualization, it is easy to follow through the lecture. Author's opinion on Parzen window in dicussion section is quite useful since it is based on the author's experience on various experiments on this method.
- (Reviewed by Sang Ho Yoon) I recommend the author to perform analysis on relationship between window width and sample sizes. The report could get benefits from testing various kernel functions.
