How to do Market Basket Analysis in Orange

แชร์
ฝัง
  • เผยแพร่เมื่อ 7 ก.ย. 2024
  • In this video we conduct Market basket analysis using the Orange software. We discuss the following metrics:
    1. Support.
    2. Lift
    3. Confidence

ความคิดเห็น • 52

  • @georgemound7098
    @georgemound7098 5 หลายเดือนก่อน +1

    Great video. It really helped me a lot. Great job

    • @theoutlier7395
      @theoutlier7395  5 หลายเดือนก่อน

      Thanks All the best.

  • @PO-bk1wv
    @PO-bk1wv 2 หลายเดือนก่อน +1

    Do we need to eliminate seasonality as we prepare data for Market Basket Analysis. For e.g., people may buy bread and butter in summer, versus the same people may buy bread and jam in winter. How would this affect the analysis? Or should I do a separate analysis for summer vs winter if I expect a strong seasonality signal? Thanks for ur views.

  • @guyvandewolfshaar5291
    @guyvandewolfshaar5291 ปีที่แล้ว

    I love you so much! thank you

  • @saalemsadeque9516
    @saalemsadeque9516 2 ปีที่แล้ว

    Very well presented.

    • @theoutlier7395
      @theoutlier7395  2 ปีที่แล้ว

      Glad you liked it!

    • @sidrakhan3709
      @sidrakhan3709 5 หลายเดือนก่อน

      Hye can you help in orange software ?

  • @jantivilea114
    @jantivilea114 ปีที่แล้ว +1

    Hi sir, thanks for your explanation. May I get the dataset?? thanks

  • @PO-bk1wv
    @PO-bk1wv 2 หลายเดือนก่อน

    At 19:40 when u expain support @18% of the total samples, you say that out of 1000 customers 18% of them purchased Freshmeat, CannedVeg, SoftDrink and Dairy. Is it true that they purchased CannedVeg despite it being a False in the row? Thanks.

    • @theoutlier7395
      @theoutlier7395  2 หลายเดือนก่อน +1

      good observation .people have not purchased cannedved

  • @eylmaz6696
    @eylmaz6696 4 หลายเดือนก่อน +1

    how can i get output as a txt or xls format

    • @theoutlier7395
      @theoutlier7395  4 หลายเดือนก่อน

      after getting the output of market basket analysis in the left hand side bottom side next to question mark you have a option to get the output in text format

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      @@theoutlier7395 can i do it for all association rules ?

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      ​@@theoutlier7395 ı want to use the datas in my thesis

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      @@theoutlier7395 i could not do it can yoy help me ? can i get output by creating automatically association rııles ?

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      @@theoutlier7395 sir i have some questions to you.
      ı have a survey executed. options are similar to this :
      1. ı know / ı dont know
      2. ı speak / ı dont speak
      so, PDF AND CDF needs to be calculated. how can I do it ? and i need to median , mod of the words as well how can i do

  • @eylmaz6696
    @eylmaz6696 7 หลายเดือนก่อน

    is this system uses Apriori algorithm ? (association rule ) ?

  • @varsharani9839
    @varsharani9839 ปีที่แล้ว

    great explaintation!!
    May I get the link to download this dataset?

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Apologies for the delayed response
      drive.google.com/file/d/1w9J1dtUT1RI0uPuJZkzIj-PaVAG3AE5X/view?usp=sharing

  • @ManaswiniAchuthan
    @ManaswiniAchuthan 9 หลายเดือนก่อน

    hello sir, the results are totally different for me than yours. i am getting more consequent = F.
    1. Is there any way to filter out and store the values that are =T in a table.
    2. why the output is coming differently sir.
    Please guide me sir. thank you. waiting for your reply.

  • @eylmaz6696
    @eylmaz6696 4 หลายเดือนก่อน

    sir i have some questions to you.
    ı have a survey executed. options are similar to this :
    1. ı know / ı dont know
    2. ı speak / ı dont speak
    so, PDF AND CDF needs to be calculated. how can I do it ? and i need to median , mod of the words as well how can i do

    • @theoutlier7395
      @theoutlier7395  4 หลายเดือนก่อน

      Not sure what is the need to calculate PDf and CDF, SPSS compte variable has what you need

    • @theoutlier7395
      @theoutlier7395  4 หลายเดือนก่อน

      Median: use SPSS Explore

    • @theoutlier7395
      @theoutlier7395  4 หลายเดือนก่อน

      Mode is not easy in SPSS

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      @@theoutlier7395 Standart deviation , mean value can be obtained via Distribiuton widget on Orange datatable ? What i obtained in the distribition widget ,s about Normal Distribution ?
      Shortly can i get results of normal distribiuton in orange ?

    • @eylmaz6696
      @eylmaz6696 4 หลายเดือนก่อน

      @@theoutlier7395 thanks sir. how can i make corelation in orange. ? i dont have numeric values

  • @ManaswiniAchuthan
    @ManaswiniAchuthan 9 หลายเดือนก่อน

    sir please give me access for the drive. i need to download this data set for assignment

  • @ayemansaqib8507
    @ayemansaqib8507 2 ปีที่แล้ว +1

    from where you downloaded the dataset

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว +1

      Sample files from IBM Modeller

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      drive.google.com/file/d/1w9J1dtUT1RI0uPuJZkzIj-PaVAG3AE5X/view?usp=sharing

  • @syazasuhaini5140
    @syazasuhaini5140 ปีที่แล้ว

    Hello sir, I want to ask question. The item that is equal to F means it was not bought, right? At 16:03,
    1. Why the antecedents still show some of the items that is equal to F? Wasnt the rules generated to show what the customers bought?
    2. Why a few of the consequent showed items that is equal to F?
    I'm a little bit confused, than you.

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Thanks for the question and good observation!

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      The item that is equal to F means it was not bought- yeah thats is right.

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      1. Ideally speaking it shouldn't display the items which haven't been bought. Many software's like IBM Modeler "strictly" displays only those products which have been bought. This is peculiar to orange where it displays some of the items where antecedents and consequents items=F.

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      2. Before executing the Apriori algorithm we can specify some constraints and ask Orange to pull out only those rules where items=T. In this exercise since i didnt put those restrictions Orange is giving all rules, even those rules where item=F

  • @swapnilanand2557
    @swapnilanand2557 ปีที่แล้ว

    Where did you get this data set

  • @eylmaz6696
    @eylmaz6696 ปีที่แล้ว

    how to construct data and how to convert data to this type

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      There are no easy answers here, We have to use Excel and a bit of pivot table to bring the data into this format.

    • @eylmaz6696
      @eylmaz6696 ปีที่แล้ว

      @@theoutlier7395 i got it. for example, if we are giving consultancy for a company, for their datasets, we need to do some processes. is it true ?

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      @@eylmaz6696 As part of data pre processing we need to create a Truth table as shown in the above dataset and then feed this in the MBA algorithm

    • @eylmaz6696
      @eylmaz6696 ปีที่แล้ว

      @@theoutlier7395 thanks sir, do big companies use orange for data analysis?

    • @eylmaz6696
      @eylmaz6696 7 หลายเดือนก่อน

      is this system uses Apriori algorithm ? (association rule ) ?@@theoutlier7395

  • @sathvikak5125
    @sathvikak5125 ปีที่แล้ว

    I am getting only F output, can you pls help me out

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Looks like yiour data has very few True cases hence u are getting only F outputs

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Please experiment by decreasing the Support, confidence try to reduce to 80 percent .Make sure the Lift is at least 1. Please dont decrease lift value lesser than one

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Let me know if it works

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      Method 2: Try ECLAT algorithm instead of apriori

  • @15.inhhieu64
    @15.inhhieu64 ปีที่แล้ว

    can you give me this data

    • @theoutlier7395
      @theoutlier7395  ปีที่แล้ว

      share ur mail id il send the data!