ML Zoomcamp Office Hours - Week #3 (by Dmitry Muzalevskiy)
Timecodes:
00:06 - Introduction and dataset options for regression homework
02:29 - Creating regression model to predict prices.
04:50 - Import warnings to reduce minor warnings.
06:55 - Data manipulation and analysis with Pandas.
08:52 - Ways to find missing values in data.
10:58 - Finding median & 15th percentile using numpy.
13:24 - Creating histograms and log transformation example.
15:36 - Using cborn library for log transformation.
17:50 - Preparing data sets for machine learning.
20:03 - Calculating RMSE for missing value imputation.
22:24 - Using np round function to calculate RMSE.
24:08 - Training linear regression with different parameters.
26:22 - Adding numbers to diagonal for stability.
28:23 - Train, predict, and print results with regularization.
30:24 - Importance of validation strategy for different models.
32:17 - Training linear regression without regularization, calculating RMSE.
34:10 - Describing results and fluctuations in values.
36:04 - Validation strategy using shuffling and RMS.
37:54 - Combining train and validation data sets.
40:01 - Calculating RMSE for regression with parameters.
42:11 - Importance of features on target variable.
44:22 - Ways to transform target variable for regression.
46:40 - Email confirmation for homework submission limitations
48:31 - Importance of feature selection and scaling.
50:27 - Scaling features important for certain solutions.
52:27 - Validation data set helps answer questions.
54:13 - Discussion on variable removal and one-hot encoding limits.
56:11 - Label encoding for machine learning models.
Links:
Course page: https://datatalks.club/courses/2021-w...
Register here: https://airtable.com/shr6Gz46UZCgJ9l6w
Course GitHub repo: https://github.com/alexeygrigorev/mlb...
Public Google calendar: https://calendar.google.com/calendar/...
The book - Machine Learning Bookcamp: http://bit.ly/mlbookcamp (Get 40% off with code "grigorevpc")
Slides: https://www.slideshare.net/AlexeyGrig...
Join DataTalks.Club: https://datatalks.club/slack.html
Our events: https://datatalks.club/events.html