## Visitors' Comments on Tutorial: LDA

We have 15 comments on this tutorial
> (date: )
By
rate up: | rate down: 2 | last rated date: 2018-06-11
Is this comment useful? |
Thank you > Thanks (date: 2010-06-29)
By Bruce Greer
I am doing some consulting using skills I have in high performance computing but in the area of statistics where I am quite weak. Your tutorial on LDA, for instance, is really helpful. Thanks much.
rate up: 138 | rate down: 80 | last rated date: 2018-06-10
Is this comment useful? |
Thank you > LDA tutorial (date: 2010-07-19)
By Abhijit Shankar Bahirat
I am grateful to author for such an wonderful feast. The text is very easy to understand and importantly briefly reviewed. Handy stuff for a novice. Thanks again, Abhijit
rate up: 121 | rate down: 76 | last rated date: 2018-06-03
Is this comment useful? |
Suggestion > statistical pattern recognition (date: 2010-08-22)
By jaisakthi
very good tutorial. NEED TUTORIAL ON PCA AND ICA
rate up: 103 | rate down: 100 | last rated date: 2018-06-03
Is this comment useful? |
Answer > ICA and PCA (date: 2010-08-24)
By Kardi
I will put those suggestion into my wish list. Thanks for your suggestion. If you like the tutorial, please recommend it to your friends
rate up: 99 | rate down: 116 | last rated date: 2018-06-07
Is this comment useful? |
Question > Discriminant analysis (date: 2010-09-17)
By Ruwanlcs
Dear Sir, I have large number of data that need to classified into two groups using discriminant analysis. I read your article about LDA and go through the your tutorial as well(http://people.revoledu.com/kardi/tutorial/LDA/Numerical%20Example.html). It was a great help me to understand the LDA. But in this example we classified data into two groups using only two features. 1. If we have more features(more than two) how we classified data into two groups ? (can you give me a simple example). 2. In my problem my data has several number of features(they represent by values) and each of data doesn\'t contain same number of features. The number of features can be changed. But I need to classify my data into two classes using discriminant analysis. Can I use LDA for that if can how can we use this. Please help me to overcome this problem if you have time. It is great if you can explain by a example because Im not a statistical guy Im only a computer since student. Thank you, Ruwan
rate up: 96 | rate down: 49 | last rated date: 2018-06-03
Is this comment useful? |
Answer > Discriminant analysis (date: 2010-09-18)
By Kardi
Hi Ruwanlcs, Number of features does not matter. Similarly using the maximization of the discriminant function, you can solve for any number of groups, not necessarily 2. If you have missing values, you probably want to use similar value or average value of the same feature, otherwise, you should use other method such as EM algorithm
rate up: 98 | rate down: 49 | last rated date: 2018-06-07
Is this comment useful? |
Others > Discriminant analysis(LDA) (date: 2010-09-27)
By Ruwanlcs
Dear Sir, I\'ve implemented a simple program for LDA and that will be classified our input objects into two grpus according to their features(two features) and training data(predefine group). But here still we have to hard coded our mean of features in each group and covariance matrix of group. >>>>>>>>> some outputs >>>>>>> Enter your object features(ex: 3.21,4.51) : 2.59,5.47 Object group is G2 (39.17382856206457,39.9518385396128) Enter your object features(ex: 21,4.51) : 2.84,5.67 Object group is G1 (46.19302206206457, 5.807523539612795) I\'d like to implement a program for object which has more than two features and which we want to classified into two groups. If you give me an example I\'ll try my best. Thank you,
rate up: 96 | rate down: 95 | last rated date: 2018-06-07
Is this comment useful? |
Others > Discriminant (date: 2015-04-16)
By Allan
This tutorial is quite good...it enabled me to understand discriminant a lot better. Now I can go into the exam to answer the questions with confidence. Thank you. Do you have others on factor analysis, SEM, profile analysis?
rate up: 139 | rate down: 82 | last rated date: 2018-06-09
Is this comment useful? |
Thank you > LDA (date: 2015-08-11)
By Riya
good one
rate up: 191 | rate down: 172 | last rated date: 2018-06-10
Is this comment useful? |
Suggestion > Bayes rule (date: 2015-10-27)
By Juraj Stevek
P(x|i) in bayes rule in LDA is not probability but probability density function. It should be stated like f(x|i).
rate up: 196 | rate down: 66 | last rated date: 2018-06-03
Is this comment useful? |
Thank you > LDA (date: 2015-11-30)
By james LaRue
The LDA tutorial was clear, I am redoing my website on bidirectional associative memory for neural networks and multichannel blind deconvolution, and I will use your style. thank you. When it comes down to it, things are not really that hard, but of course, some prefer the veil of mystery. Anyone with a good grasp of pythagorean's theorem and trigonometry can go along way.
rate up: 179 | rate down: 66 | last rated date: 2018-06-03
Is this comment useful? |
Suggestion > Data Miming (date: 2016-03-14)
Hello Dr teknomo, I like your articles and use them in my data mining class. If you can post things in pdf files so that formulas can be eeasier read it would be nice. Thanks Dimitrios
rate up: 138 | rate down: 124 | last rated date: 2018-06-11
Is this comment useful? |
Thank you > LDA Example (date: 2016-03-24)
By Abhishek Aravind Kulkarni
This tutorial has one of the better numerical examples of LDA solutions on the Web.
rate up: 82 | rate down: 45 | last rated date: 2018-06-06
Is this comment useful? |
Thank you > Excellent Tutorial (date: 2017-03-31)
Thank you Dr. Kardi for your excellent tutorial. I hope you give us access to some of your restricted materials as well. Again, thank you.
rate up: 94 | rate down: 40 | last rated date: 2018-06-03
Is this comment useful? |

Please drop me a note if you find this tutorial is useful. I also welcome any suggestions to improve it. Let me know what worked for you and what didn't by answering the feedback form below. Sorry, I cannot help you to solve your assignment or homework. Check the FAQ before you post your question.

```Rate the quality    Excellent  Good  Fair  Poor  Bad

Why do you rated this tutorial this way? (optional)