Support Vector Machines (SVM) Overview and Demo using R

Поділитися
Вставка
  • Опубліковано 29 лис 2024

КОМЕНТАРІ • 50

  • @booradleyboo
    @booradleyboo 7 років тому +78

    Here is the script used in this video for anyone who doesn't have time to type it out:
    library("e1071")
    plot(iris)
    plot(iris$Sepal.Length, iris$Sepal.Width, col=iris$Species)
    plot(iris$Petal.Length, iris$Petal.Width, col=iris$Species)
    s

    • @nkundukozerajanvier162
      @nkundukozerajanvier162 7 років тому +1

      thank you for your time taken

    • @lifewith7dogs106
      @lifewith7dogs106 5 років тому

      THANKSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS

    • @saikalyangonuguntla594
      @saikalyangonuguntla594 4 роки тому

      tq

    • @booradleyboo
      @booradleyboo 2 роки тому +1

      @@Jaime-sf8ep run the following either at the top of the script (one time only) or directly in the console:
      install.packages("e1071")

  • @kalyanasundaramsp8267
    @kalyanasundaramsp8267 6 років тому +3

    amazing melvin..what a free flow and for lay man like me ..its easy to understand

  • @ZxJorge
    @ZxJorge 4 роки тому +1

    Tnk man. You saved me. Grettings from Mexico!

  • @andialpacino8331
    @andialpacino8331 5 років тому +2

    thank you very much, this video very cleary explanation dan very helpful bro
    Keep going
    Regards from Indonesia.

  • @isaiahhernandez1678
    @isaiahhernandez1678 4 роки тому +1

    GREAT video. Thank you so much

  • @josueperez4333
    @josueperez4333 4 роки тому +1

    “Introduction to Machine Learning With R” O’Reilly books are pretty good

  • @kobe8poprox
    @kobe8poprox 7 років тому +2

    Excellent Video! Very well explained! Could you maybe do another one for One-Class SVM?

  • @paulchabert5340
    @paulchabert5340 6 років тому +1

    Really nice, simple and useful code! thank you

  • @krishnakanth4718
    @krishnakanth4718 8 років тому

    I really the way the explanation was , it was very good .

  • @shreyaschaturvedi1933
    @shreyaschaturvedi1933 4 роки тому

    can someone explain the plot step to me? i'm confused on what iris_train[,col] means

  • @nathanyakich3152
    @nathanyakich3152 8 років тому

    This video was a great help! Thank you!
    when I try and view the data in a table instead of in a histogram (near the end of the video), I get an error that "all arguments must be the same length"

  • @juleswombat5309
    @juleswombat5309 8 років тому

    Impressive demo

  • @erigits
    @erigits 8 років тому

    Good work, does the features (no of features) affect the optimal hyperplane

  • @mq.1554
    @mq.1554 7 років тому

    You have covered pretty much everything in this video but I guess at some points I need some explanation. If u can please answer my questions.
    1) On what basis you actually defined the range values for the cost variable?
    2) I was reading their practical guide and I found that 'Gamma' is also as important as 'Cost' but u never discussed anything.
    3) I know that at the end we are calculating the mean value which actually representing the system accuracy. Can you tell me what this 'number of vectors' is representing? Like i am solving a two spiral problem so i am getting 130 no. of vectors. I am curious cuz i am just passing 132 rows of data. Is that bad or good ? or how do you interpret?
    I would be very thankful to you if you can answer my questions.

  • @shareefamohamed193
    @shareefamohamed193 4 роки тому

    Sir, how to identify the important variables in SVM when we have a set of variables?

  • @christineanietert179
    @christineanietert179 8 років тому +3

    When I tried to make "iris_train

    • @MrRyuuzakiiii
      @MrRyuuzakiiii 8 років тому

      try this code:
      library(e1071)
      plot(iris$Petal.Width, iris$Petal.Length, col=iris$Species)
      iris_train

    • @kishumeshu
      @kishumeshu 7 років тому

      Please, how can we install the e1071 package ?

    • @付泽-w1y
      @付泽-w1y 7 років тому

      you can input: install.packages("e1071")

  • @hariramrajan6901
    @hariramrajan6901 6 років тому

    This seems to be inspired from Patrick Winston's lecture on SVM from MIT. Great work nevertheless :)

  • @a.a3265
    @a.a3265 8 років тому

    could you please explain svm with a simple example by using data for binary classification

  • @apoorvspydy
    @apoorvspydy 7 років тому +1

    Nice explanation :)

  • @rawiasammout5833
    @rawiasammout5833 8 років тому

    Thanks really for this video please could you give us an example of non linear svm classification and thank you

  • @dandi6098
    @dandi6098 6 років тому

    i plot my model, but it doesnt display the plot.
    plot(modelsvm, dataTrain[,c])
    ^ modelsvm is same as your svmfit and dataTrain is same as your iris_train and c is same as your col
    how did i fix this ?

    • @caroline132456
      @caroline132456 5 років тому

      Your variable corresponding to "Species" in the iris dataset has to be in the format "factor" ;)

  • @reidhibbard963
    @reidhibbard963 6 років тому

    attempting to do this on a different data set but cannot identify two classifiers that appear to be linearly separable, any suggestions?

  • @vivekmangipudi3773
    @vivekmangipudi3773 8 років тому +1

    good video. is there a github link or website with the code?

  • @dipayansarkar3594
    @dipayansarkar3594 8 років тому +1

    Can we use SVM for time series forecasting? I do not have any feature set other than year, week & sales (in units)...

    • @melvinl5797
      @melvinl5797  8 років тому +1

      Have not tried using SVM for time series forecasting however seems
      doable stackoverflow.com/questions/31048964/time-series-forecasting-using-support-vector-machine

  • @me3jab1
    @me3jab1 5 років тому

    how many karnels type that we have for SVM ?

  • @mmaandbjjlondon5238
    @mmaandbjjlondon5238 3 роки тому

    Sorry to ask, do you offer private lessons?thanks

    • @melvinl5797
      @melvinl5797  3 роки тому

      Thanks for asking but I don’t do any private lessons

  • @cherryberry5997
    @cherryberry5997 8 років тому

    could you please help me, I have a class with 8 levels. how would I use svm to create the 8 levels ???

  • @samjoshua192
    @samjoshua192 8 років тому

    HI all, I new to R, can some explain why iris_test=iris[-s, col] returns 50 observations?

  • @motasemha3404
    @motasemha3404 8 років тому

    great thank u.
    with which data or when we can use SVM ? plz
    is good with numeric data?
    should we convert our numeric data to factors?.....

  • @srisivanicharan6079
    @srisivanicharan6079 8 років тому

    When i tried to execute this piece of code iris_train

  • @matteotesti2337
    @matteotesti2337 8 років тому

    Hello there I used R studio but I got this error:
    Package 'e1701' was built under version 3.2.5
    How can I update my R studio?
    Many thanks

    • @nathanyakich3152
      @nathanyakich3152 8 років тому

      There's a package called "installr" that can update you to the latest version

  • @ajumborogueprince7685
    @ajumborogueprince7685 6 років тому +1

    Why aren't you scaling the data?

  • @ashishjain6119
    @ashishjain6119 7 років тому

    too fast. could not get use of cost and tuned properly.

  • @Paganel75
    @Paganel75 7 років тому +2

    "Er... Er... Er... Er... Er... Er..." How _fascinating_ !

  • @Jaime-sf8ep
    @Jaime-sf8ep 2 роки тому

    how do you install the SVM package