Timeline for Encoding for classifiers
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 7, 2020 at 12:12 | vote | accept | Math | ||
| Aug 29, 2020 at 9:09 | comment | added | hssay | I added a detail in the answer. For joining, you need to figure out if you are going to first aggregate the data at user level (which I think you should). In that case, the joining key will be user ID. | |
| Aug 29, 2020 at 9:08 | history | edited | hssay | CC BY-SA 4.0 |
added 357 characters in body
|
| Aug 28, 2020 at 14:07 | comment | added | Math | Please correct me if I am wrong: first I work on Text column. I apply text vectorizer to this column and I will get a dataframe (A) having the same columns as my original one, except for Text column, which will contain some numerical value (I suppose). Then, I consider again the original dataframe, but workingn only on symbol and note: then, I apply one hot encoder to them and I consider this new dataframe as B. Accounts are not unique. How should I join them? I could create an ID in both, but I think I could join on more terms:for example on Account and age.One account may have multiple texts | |
| Aug 28, 2020 at 14:03 | comment | added | Math | Thank you so much hssay for your answer. May I ask you to do an example for Text column and Symbols? It has not completely clear to me after using encoding and text vectorizer how my data would look like. Also when you say two different data frames. | |
| Aug 28, 2020 at 12:05 | history | answered | hssay | CC BY-SA 4.0 |