@@ -21,4 +21,23 @@ The dataset, found in the `data/` folder, is scraped from user pages of [nomadli
## Working with the datasets: See boilerplate code in src/Starter.ipynb
- 3. run `jupyter notebook`
- 4. select `src/Starter.ipynb`
\ No newline at end of file
- 4. select `src/Starter.ipynb`
-
# Progress
## To be done on 25.11. (see [Starter](https://gitlab.fi.muni.cz/xslanin/pv254-city-recommender/blob/master/src/Starter.ipynb)):
- assign responsibilities for:
- - come up with more approaches for recommendations using similarities (see [Approach 1 in Starter](https://gitlab.fi.muni.cz/xslanin/pv254-city-recommender/blob/master/src/Starter.ipynb))
- - dimensionality reduction / visualization of city distances using PCA/t-SNE
- - data analysis on cities dataset (partially started by Martin)
- - come up with more features for cities, e.g. sport, art, nature
- - create 1 blackbox recommendation solution for comparison purposes
- - think about testing scenarios, focus on use cases
- - - e.g. "I've been to (generic city)", "I've been to (list of generic cities)", "I've been to (generic city and obscure one)", etc.
- - - e.g. "I care about (feature)", "I care about (list of similar features)", "I care about (list of disimilar features)", etc.
- - - combination: e.g. "I care about (feature) and i've been to (list of generic cities)"
- - - conditioning on a location: e.g. "I care about (feature) and i've been to (list of generic and obscure cities) and i want to go to (continent)"
- - testing: come up with a dataset of testing users based on the use cases
- - run all the algorithms on it and come up with explanations (if there are)
