Machine Learning | Fuzzy C Means
Вставка
- Опубліковано 28 сер 2024
- This algorithm works by assigning membership to each data point corresponding to each cluster centre based on the distance between the cluster centre and the data point. More the data is near to the cluster centre more is its
membership towards the particular cluster centre. Clearly, the summation of membership of each data point should be
equal to one. #MachineLearning #FuzzyCMeans #FCM
Machine Learning 👉 • Machine Learning
Artificial Intelligence 👉 • Artificial Intelligenc...
Cloud Computing 👉 • Cloud Computing Tutorials
Wireless Technology 👉 • Wireless Technology Tu...
Data Mining 👉 • Data Mining & Business...
Simulation Modeling 👉 • Simulation Modeling Tu...
Big Data 👉 • Big Data Anaytics
Blockchain 👉 • Blockchain Technology
IOT 👉 • Internet Of Things
Follow me on Instagram 👉 / ngnieredteacher
Visit my Profile 👉 / reng99
Support my work on Patreon 👉 / ranjiraj
For notes👉 github.com/ranjiGT/ML-latex-amendments
Welcome
I'm Adnan, a university student
I would like you to help me with programming about the FUZZY C MEANS algorithm
Thank you
It became too much theory. Add a numerical example explaining the flow, which will clear all the doubts.
Few comments from me, feel free to comment on it.
1. I believe the 2nd summation term at time 12:44 needs to be corrected. The summation for a given cluster across all members is only lower bounded (greater than zero). To give an example, if most of the members are close to one cluster, that would result in a summation a lot higher than 1. You can also refer to "Membership functions in the fuzzy C-means algorithm" for the same
2. At time 16:10 I think you are mixing the usage of 'set' and 'cluster'. mu_i,j will be equal to 1 if the member i belongs to the cluster j and exactly overlaps with the centroid or cluster center.
3. At time 16:10, I am not quite sure I follow how you decided the A matrix to be 3x3 size. Rather it would be easier if we followed distance squared logic to explain the final expression instead of using a positive definite matrix approach.
4. At time 18:55, At the very first time of running the algorithm, how exactly do you start with U before the membership function is identified? I assume you randomly assign values to the membership function.
5. At time 21:50, while describing the condition for repeating in loops, I believe as long as the cluster centroid (or mean of Ai) hasn't shifted from the last step, we have converged as there will be no further changes to the centroids.
Not able to get clarity bro!
Thanks Ranji Raj.Very informative.
fuzzy interval is between 0
Informative. Thank you so much.
Thank you so much sir
All the best
@@RanjiRaj18 overlapping cluster means the same data points are available in more than two clusters . Is it sir?
Yes, you are correct.
@@RanjiRaj18 you have mentioned dik for calculating fuzzy membership. i is a data point. j is the clusters. What is k? Kindly reply it sir?
Great video, very informative.
What a good video, thank you Sir and hello from México
It was very helpful thnx
Most welcome 😊
Please try to be careful while writing equations on the board. While writing the fuzzy membership function’s definition you wrote like this … 1 ≤ µij ≤ 0
Thank you so much.
Never explain things with too much theory.
Exactly
Don't skip the steps and need some clarification.
How a number can be greater than 1 but less than 0 i.e. membership function?
Sir, if i had a different scale value of my attributes ( one attribute is binary value, one attribute is 0 - 10000, one attribute is 0 - 1000 ) then, how to normalize my attribute?
Should 'r' be strictly greater than 1?
Yes
hi sir, do you have code for the algorithm?
Yes I have. You can get here: github.com/ranjiGT/Python-Hackerrank/blob/main/Fuzzy-c-Means.ipynb
good informative video
Passa seu whats app ai irmão