Evaluate the results in various aspects

Prepare the dataset as introduced before:

[1]:

import rsdiv as rs

loader = rs.MovieLens1MDownLoader()
ratings = loader.read_ratings()
items = loader.read_items()

Various diversity metrics

Load the evaluator to analyze the results, say, the Gini coefficient metric:

[2]:

metrics = rs.DiversityMetrics()
metrics.gini_coefficient(ratings['movieId'])

[2]:

0.6335616301416965

The nested input type (List[List[str]]-like) is also favorable. This is especially useful to evaluate the diversity on the topic scale:

[3]:

metrics.gini_coefficient(items['genres'])

[3]:

0.5158655846858095

Shannon Index and Effective Catalog Size are also available with the same usage.

Show the distribution of a given data source

[4]:

distribution = metrics.get_distribution(items['genres'])
../_images/notebooks_evaluate-the-results-in-various-aspects_10_0.png

Draw a Lorenz curve graph for insights

Lorenz curve is a graphical representation of the distribution, the cumulative proportion of species is plotted against the cumulative proportion of individuals. This feature is also supported by rsdiv for helping practitioners’ analysis.

[5]:

metrics.get_lorenz_curve(ratings['movieId'])
../_images/notebooks_evaluate-the-results-in-various-aspects_13_0.png

Show the distribution of a given data source

The unbalance of the data distribution can be well illustrated by both barplot and sorted DataFrame:

[6]:

distribution = metrics.get_distribution(items['genres'])
../_images/notebooks_evaluate-the-results-in-various-aspects_16_0.png 
[7]:

distribution

[7]:
category count percentage
0 Drama 1603 0.250156
1 Comedy 1200 0.187266
2 Action 503 0.078496
3 Thriller 492 0.076779
4 Romance 471 0.073502
5 Horror 343 0.053527
6 Adventure 283 0.044164
7 Sci-Fi 276 0.043071
8 Children's 251 0.039170
9 Crime 211 0.032928
10 War 143 0.022316
11 Documentary 127 0.019819
12 Musical 114 0.017790
13 Mystery 106 0.016542
14 Animation 105 0.016386
15 Fantasy 68 0.010612
16 Western 68 0.010612
17 Film-Noir 44 0.006866

Evaluate the unbalance from a sense of location

rsdiv provides the encoders including geography encoding function to improve the intuitive understanding for practitioners, to start with the random values:

[8]:

import numpy as np

geo = rs.GeoEncoder()
df, gdict = geo.read_source()
rng = np.random.default_rng(42)
df['random_values'] = rng.random(len(df))

[9]:

geo.draw_geo_graph(df, 'random_values', 'name')

Data type cannot be displayed: application/vnd.plotly.v1+json

image0