How does one come up with the graph? is it based on data sets that was collected over time, if within an exam; would the graph be provided to us? or is there a way to draw out the graph based on the rules and crisp values that was given to us?
The graph, the linguistics value, the scale of the membership as well as the the graph shapes are from experts. This is a continuation from 5/6 prev subtopics which started with acquire knowledge from human expert. Of coz, we can also get them by studying recorded data. Hope it helps
Hi, we certainly will have to go based on experts hunches for the value or percentage. Else, knowledge engineer can come with suggestion and work with the expert for validation.
could not figure out how would you come up to the slow and fast line based on 10% slow and 75% fast? can you give more detail on how to draw these lines? Thanks
at minute 1:44 for example; Temperature given in the question was 65 deg F. So all we have to do is draw a vertical line on X-axis on 65F and read the value of its corresponding Y-axis that crossed any of the linguistics var (blue, cyan, yellow and red). in this case it crossed over with yellow with value 0.7 (estimated as not in graph/grid paper) and cyan at 0.1(again, estimated as not in graph/grid paper). The 10% and 70% was from the rules defuzzification where you need to choose min(for AND) and max (for OR). hope it helps. (so sorry for a very late reply)
Hi Jean, thanks for watching. This video is actually a continuation of tutorial discussion in class. I have been meaning to create a better video but couldn't find a good time yet. I'll try to explain here. Four main steps in fuzzy logic:
1. Fuzzification. we need to converts values into linguistic variables and assigned membership values.Where do we get the values? Either from domain expert, or statistical analysis. These values may change from time to time too. An intellligent FLS idea it a dynamic change of these range of membership function based on training data.
2. Construct the fuzzy rules. the rules are similar to Decision tree rules where we have IF, OR, AND, THEN.. but, with the linguistic variables in place.
3. Inference. This is where we try to map the fuzzy rules into the membership graphs of all our parameter (the precedents on the header part of our rules) to the tail part. We do this to all of our fuzzy rules. The membership function shapes, depends on the degree of membership.
4. Defuzzification. This is basically to make the decisions (usually overlapped /conflicting decision from the inference process(eg in min 5.18 of my video, we have sugggestion to drive slow[0.1] ad fast[0.7]), into a crisp output (as in min 8). there are may method however to do this. the video used centroid and weighted mean. we can also use mean-max (middle of max), weighted average, center of gravity etc.. try read this: cse.iitkgp.ac.in/~dsamanta/courses/archive/sca/Archives/Chapter%205%20Defuzzification%20Methods.pdf if it helps
السلام عليكم Firstly, hopefully you are doing well. Secondly, I have implemented this example in matlab but when I put 65 temp and 25 for cloud I didn’t get the speed that you have calculated which is 68.75. So could you please explain what is the problem ? Thank you
Hi! I'm working on a thesis with fuzzy logic and I was wondering is it possible for me to design my own rules based on certain information? I would like to contact you please. Regards.
Same here. Im working on Electronic Nose Food Freshness with Fuzzy logic. Basically I'm using fuzzy to predict/determine food freshness. Based on the detection of gas emitted from food by gas sensor connected with arduino
The equations are wrong. To demonstrate the cool point is 0.25 not 0.3 if you round up like you did the answer is going to be a totally different answer than reality.
![Twenty One Pilots - “The Line” (from Arcane Season 2) [Official Music Video]](http://i.ytimg.com/vi/E2Rj2gQAyPA/mqdefault.jpg)
Thank you for your good explanation about the topic. I believe it will help me to clarify more about the topic. Love from india🇮🇳
hello
I couldn't understand why did you choose 25 and 75 for 10% slow and 70% fast respectively?
25% is where slow is at 100% and 75 is where fast is at 100%.
Thank you, it was a really clear explanation!
*Your voice is very beautiful ❤️😍*
How does one come up with the graph? is it based on data sets that was collected over time, if within an exam; would the graph be provided to us? or is there a way to draw out the graph based on the rules and crisp values that was given to us?
The graph, the linguistics value, the scale of the membership as well as the the graph shapes are from experts. This is a continuation from 5/6 prev subtopics which started with acquire knowledge from human expert.
Of coz, we can also get them by studying recorded data. Hope it helps
Thank you so much for the clarification and reply! @@afzanadam5076
PER ITALIANI: Ho creato l'unica playlist su RUclips Italia sulla logica fuzzy! la trovate quì -> ruclips.net/p/PL-tCoHPn6YlafYUe6bSnpZJe2MSaHer8h
how do we calculate the y-axis value forcool(0,3) and warm(0,7)?
Hi, Is there a specific rule to define fuzzy values or it just depends on experts' domain information?
Hi, we certainly will have to go based on experts hunches for the value or percentage. Else, knowledge engineer can come with suggestion and work with the expert for validation.
@@afzanadam5076 Just like I assumed, but needed some confirmation. Thanks for the quick response...
Great video. Nice explaination
could not figure out how would you come up to the slow and fast line based on 10% slow and 75% fast? can you give more detail on how to draw these lines? Thanks
at minute 1:44 for example; Temperature given in the question was 65 deg F. So all we have to do is draw a vertical line on X-axis on 65F and read the value of its corresponding Y-axis that crossed any of the linguistics var (blue, cyan, yellow and red). in this case it crossed over with yellow with value 0.7 (estimated as not in graph/grid paper) and cyan at 0.1(again, estimated as not in graph/grid paper).
The 10% and 70% was from the rules defuzzification where you need to choose min(for AND) and max (for OR).
hope it helps. (so sorry for a very late reply)
@@afzanadam5076 but if you do so, it will be 0.7 and 0.3
that's what you have written. Not getting how to do this
@@rajapandey773 pls rewatch. 70 & 30 is for rule on temperature, while 70 and 10 is on rule for cloud cover.
what if the cloud cover is 50? do we use complement?
Quite helpful video to understand the basic idea of fuzzy logic. The link which you said is missing however.
Thank you @Afzan_Adam ......Mohamed from Egypt.
Thank you very much ! Couldn’t see the link you told in the end of video , if you can send it to me in canada.youssef@gmail.com. Thank you very much
very helpful thanks
Hi, this is a good knowledge about fuzzy logic, one question about is how to changes the values range of the membership functions.
Hi Jean, thanks for watching. This video is actually a continuation of tutorial discussion in class. I have been meaning to create a better video but couldn't find a good time yet. I'll try to explain here. Four main steps in fuzzy logic:
1. Fuzzification. we need to converts values into linguistic variables and assigned membership values.Where do we get the values? Either from domain expert, or statistical analysis. These values may change from time to time too. An intellligent FLS idea it a dynamic change of these range of membership function based on training data.
2. Construct the fuzzy rules. the rules are similar to Decision tree rules where we have IF, OR, AND, THEN.. but, with the linguistic variables in place.
3. Inference. This is where we try to map the fuzzy rules into the membership graphs of all our parameter (the precedents on the header part of our rules) to the tail part. We do this to all of our fuzzy rules. The membership function shapes, depends on the degree of membership.
4. Defuzzification. This is basically to make the decisions (usually overlapped /conflicting decision from the inference process(eg in min 5.18 of my video, we have sugggestion to drive slow[0.1] ad fast[0.7]), into a crisp output (as in min 8). there are may method however to do this. the video used centroid and weighted mean. we can also use mean-max (middle of max), weighted average, center of gravity etc.. try read this: cse.iitkgp.ac.in/~dsamanta/courses/archive/sca/Archives/Chapter%205%20Defuzzification%20Methods.pdf if it helps
Great!
Simple and nice explanation
Thanks
Thankyou , it was nice explaination.
Very good
Ty. Gorgeous lady!!!
I need an membership function which enables me to get the same input at the out put
the best video
So useful thank you♥️
its not helpful even a little bit, even you don't tell us how to draw output lines :/
Hello I have HW and I need help can I send it to you?
Thank you
is it a bachelor or master lesson?
Cool thanks
your pronunciation is not clear
adam hintli amk ne bekliyonuz, bu aksanlara bu telaffuzlara alisamadiysaniz bu meslegi birakin
clear
very concise !
السلام عليكم
Firstly, hopefully you are doing well. Secondly, I have implemented this example in matlab but when I put 65 temp and 25 for cloud I didn’t get the speed that you have calculated which is 68.75. So could you please explain what is the problem ?
Thank you
I cant really comment on that as one of the reasons probably, could be the code, assigning of values or the defuzzification method.
Hi! I'm working on a thesis with fuzzy logic and I was wondering is it possible for me to design my own rules based on certain information? I would like to contact you please. Regards.
email me
@@afzanadam5076 Thanks! I'll send you an email ASAP.
Same here. Im working on Electronic Nose Food Freshness with Fuzzy logic. Basically I'm using fuzzy to predict/determine food freshness. Based on the detection of gas emitted from food by gas sensor connected with arduino
thank you so much :)) i was fed up with this shit
The equations are wrong.
To demonstrate the cool point is 0.25 not 0.3 if you round up like you did the answer is going to be a totally different answer than reality.
I need a
At 7:20 the percentage of Slow randomly changes from 10% to 20%