import pandas as pd
import seaborn as sns
snap = pd.read_csv("test_results.csv")
df.sample(4)
method | traintime | F1 | F1_test | MI | RAND | F-M | params | graph | |
---|---|---|---|---|---|---|---|---|---|
44 | Node2Vec | 300.006000 | 0.850000 | 0.505000 | 0.644000 | 0.520000 | 0.544000 | {'epochs': 80, 'keep_walks': False, 'n_compone... | snap |
19 | SKLearnEmbedder | 7.671000 | 0.879000 | 0.469000 | 0.665000 | 0.589000 | 0.608000 | {'embedder__a': None, 'embedder__angular_rp_fo... | snap |
63 | Glove | 16.967000 | 0.622000 | 0.122000 | 0.001000 | -0.000000 | 0.221000 | {'learning_rate': 0.1, 'max_epoch': 50000, 'ma... | blogcatalog |
87 | SKLearnEmbedder | 464.722672 | 0.545764 | 0.066072 | 0.046016 | 0.010172 | 0.071656 | {'embedder__a': None, 'embedder__angular_rp_fo... | blogcatalog |
nv = df.loc[df.method == 'Node2Vec'].reset_index(drop=True)
nv = nv.join(pd.DataFrame.from_records(nv.params.apply(eval).values))
nv.sample(3)
method | traintime | F1 | F1_test | MI | RAND | F-M | params | graph | epochs | keep_walks | n_components | neighbor_weight | return_weight | threads | w2vparams | walklen | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
4 | Node2Vec | 74.403 | 0.868 | 0.436 | 0.660 | 0.540 | 0.563 | {'epochs': 20, 'keep_walks': False, 'n_compone... | snap | 20 | False | None | 0.3 | 3.0 | 4 | {'window': 10, 'negative': 5, 'iter': 5, 'ns_e... | 80 |
20 | Node2Vec | 300.063 | 0.871 | 0.430 | 0.659 | 0.539 | 0.563 | {'epochs': 80, 'keep_walks': False, 'n_compone... | snap | 80 | False | None | 0.3 | 1.0 | 4 | {'window': 10, 'negative': 5, 'iter': 5, 'ns_e... | 80 |
35 | Node2Vec | 334.563 | 0.894 | 0.481 | 0.622 | 0.480 | 0.514 | {'epochs': 80, 'keep_walks': False, 'n_compone... | snap | 80 | False | None | 3.0 | 3.0 | 4 | {'window': 10, 'negative': 5, 'iter': 5, 'ns_e... | 80 |