Machine Learning | Kalicharan

Discussion in 'Big Data and Analytics' started by Nishant_Singh, Jun 1, 2019.

  1. Nishant_Singh

    Nishant_Singh Well-Known Member
    Simplilearn Support

    Joined:
    Aug 1, 2018
    Messages:
    222
    Likes Received:
    31
    #1
  2. NIRUPOM DAS

    NIRUPOM DAS Member

    Joined:
    Mar 11, 2019
    Messages:
    3
    Likes Received:
    0
    Hi Kalicharan, as you were saying can you please share the links of some good videos and books on machine Learning and also the homework of class on 2nd June that we need to finish by next class ?
    I have checked the shared google drive, but its empty.
     
    #2
  3. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
    Yes, the last exercise you will need to finish by next class and let me know how it went
     
    #3
  4. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
    Please see the copy of books that i read to understand AI and machine learning
     

    Attached Files:

    #4
    Sandeep Subbenaik likes this.
  5. NIRUPOM DAS

    NIRUPOM DAS Member

    Joined:
    Mar 11, 2019
    Messages:
    3
    Likes Received:
    0
    Hi Kalicharan,

    1. Can you please share one data set where we should use Imputer function and then we can solve some business objective. Need the problem statement also, which we should solve.

    2. From lesson 4 : PCA, LDA.
    We have done PCA for breast cancer problem, as there was 30 dimensions and we made it to 2 and then put logistic regression. So can we generalize, that if we have lets say more than 10 dimensions, we ALWAYS should use dimensionality reduction, so that our algorithm will work properly OR there is any catch here ?

    3. Is it possible to to check and print the column name which has unique values or have categorical data ? Wanted to know which ALL teh columns have unique values using code and not going through the CSV file manually.

    4. I have submitted 1 project. There are some tough weeks coming for me (professionally), so will really appreciate if you check my submitted project and let me know, if I have done it properly or not?

    Regards,
    Nirupom
     
    #5
    Last edited: Jun 25, 2019
  6. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
    1. Use any of your data like the diabetes data or the Mtcars data and delete some one the data . Use Imputer to substitute the data and see the difference in performance.
    2. It is always better to reduce the dimensionality.such that visually you can see the performance of the algorithm. If your algorithm already gets a good performance then you dont need to reduce the dimensionality
    3. Yes, please see an example https://chrisalbon.com/python/data_wrangling/pandas_list_unique_values_in_column/
    4. Sure, let me have a look
     
    #6
  7. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  8. NIRUPOM DAS

    NIRUPOM DAS Member

    Joined:
    Mar 11, 2019
    Messages:
    3
    Likes Received:
    0
    for question 3, what I meant was if you have 30 columns, and within that lets say 5 columns have categorical data. So what I need is to print those 5 column name.
    df.<column name>.unique() : give us the unique value for that particular column. I need to print all of the 5 column names.... I do not want to go one by one column and check for unique value.
     
    #8
  9. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  10. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  11. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  12. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  13. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  14. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
  15. Sudhir Kumar_10

    Joined:
    Feb 24, 2019
    Messages:
    4
    Likes Received:
    1
    Hello Kalicharan and support team

    The recording for 29th June for machine learning course seems to be corrupt.Its only 95MB in size. and while opening its showing wrong format of file.
     

    Attached Files:

    #15
  16. Sudhir Kumar_10

    Joined:
    Feb 24, 2019
    Messages:
    4
    Likes Received:
    1
    Hello support team

    The recording for 29th June for machine learning course seems to be corrupt.Its only 95MB in size. and while opening its showing wrong format of file.
     

    Attached Files:

    #16
  17. Sandeep Subbenaik

    Sandeep Subbenaik New Member

    Joined:
    May 29, 2019
    Messages:
    1
    Likes Received:
    0
    Hello team, please send me recording links. I am not able to see any recordings on lms since i was hooked up to this course via Backend.
     
    #17
  18. Kalicharan_1

    Kalicharan_1 Member

    Joined:
    Apr 8, 2019
    Messages:
    11
    Likes Received:
    1
    This link should help
    https://stackoverflow.com/questions/29803093/check-which-columns-in-dataframe-are-categorical
     
    #18
  19. Sriram Srinivasaraghavan

    Joined:
    May 8, 2019
    Messages:
    1
    Likes Received:
    0
    Hi Kali,

    How to deal with a feature that is of Date, in a dataset for Classification.
    I thought of label encoding the categorical columns to plot a heatmap/corr.
    Before that data has to be LabelEncoded. But I am not sure if it will make sense to label encode date values.
    At the same, the date value can't be omitted just like that.
    Data set is about the events happened in the past 10years along with the date. So I feel date might have some significance here.
    Same date around every year might have some correlation.

    Kindly advise how to handle the date here.

    Regards,
    Sriram S
     
    #19

Share This Page