Hiya and welcome everybody to the ultimate a part of our challenge, Challenge 3. On this session, we’ll be utilizing the instruments we’ve already realized about to construct a mannequin and discover the query of ethics in decoding and utilizing the mannequin in the true world.
For at the moment, our foremost objectives are:
- Put together a family dataset for binary classification, particularly specializing in the ‘household_demographics’ desk within the database.
- Group excessive cardinality categorical options right into a smaller variety of classes for exploration.
- Create a logistic regression mannequin to foretell extreme injury and clarify mannequin predictions utilizing odds ratios.
- Examine vital options associated to “caste,” a demographic piece within the database, to know how the mannequin makes use of these options.
Following our ML workflow, we’ll work with the brand new desk, the household_demographics desk, for information preparation. Our exploration will contain grouping excessive cardinality categorical options and the break up will stay the identical, simply prepare and take a look at break up this time. Constructing the mannequin will likely be acquainted, serving as one other alternative to construct muscle reminiscence. In terms of new content material, we’ll be speaking outcomes by taking a look at odds ratios and investigating the obvious…