Download the dataset and place it in data directory.
This project uses DVC. To see results of several tries do:
$ dvc metrics show -a -TBelow you find a sample output of this command:
working tree:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
2408,
735
],
[
584,
2523
]
],
"classification_report": {
"0": {
"precision": 0.8048128342245989,
"recall": 0.7661469933184856,
"f1-score": 0.7850040749796251,
"support": 3143
},
"1": {
"precision": 0.7744014732965009,
"recall": 0.8120373350498874,
"f1-score": 0.7927729772191673,
"support": 3107
},
"accuracy": 0.78896,
"macro avg": {
"precision": 0.7896071537605499,
"recall": 0.7890921641841865,
"f1-score": 0.7888885260993962,
"support": 6250
},
"weighted avg": {
"precision": 0.7896947384800229,
"recall": 0.78896,
"f1-score": 0.7888661516609463,
"support": 6250
}
},
"accuracy_score": 0.78896
}
}
master:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
2408,
735
],
[
584,
2523
]
],
"classification_report": {
"0": {
"precision": 0.8048128342245989,
"recall": 0.7661469933184856,
"f1-score": 0.7850040749796251,
"support": 3143
},
"1": {
"precision": 0.7744014732965009,
"recall": 0.8120373350498874,
"f1-score": 0.7927729772191673,
"support": 3107
},
"accuracy": 0.78896,
"macro avg": {
"precision": 0.7896071537605499,
"recall": 0.7890921641841865,
"f1-score": 0.7888885260993962,
"support": 6250
},
"weighted avg": {
"precision": 0.7896947384800229,
"recall": 0.78896,
"f1-score": 0.7888661516609463,
"support": 6250
}
},
"accuracy_score": 0.78896
}
}
exp/baseline:
model/metrics.json:
{
"confusion_matrix": [
[
7883,
4617
],
[
8007,
4493
]
],
"classification_report": {
"0": {
"precision": 0.4960981749528005,
"recall": 0.63064,
"f1-score": 0.5553363860514265,
"support": 12500
},
"1": {
"precision": 0.4931942919868277,
"recall": 0.35944,
"f1-score": 0.41582600647848217,
"support": 12500
},
"accuracy": 0.49504,
"macro avg": {
"precision": 0.49464623346981407,
"recall": 0.49504,
"f1-score": 0.48558119626495433,
"support": 25000
},
"weighted avg": {
"precision": 0.494646233469814,
"recall": 0.49504,
"f1-score": 0.48558119626495433,
"support": 25000
}
},
"accuracy_score": 0.49504
}
exp/logistic-regression:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
8415,
4085
],
[
4190,
8310
]
],
"classification_report": {
"0": {
"precision": 0.6675922253074177,
"recall": 0.6732,
"f1-score": 0.6703843855805617,
"support": 12500
},
"1": {
"precision": 0.6704316256555063,
"recall": 0.6648,
"f1-score": 0.6676039365334404,
"support": 12500
},
"accuracy": 0.669,
"macro avg": {
"precision": 0.6690119254814619,
"recall": 0.669,
"f1-score": 0.668994161057001,
"support": 25000
},
"weighted avg": {
"precision": 0.669011925481462,
"recall": 0.669,
"f1-score": 0.668994161057001,
"support": 25000
}
},
"accuracy_score": 0.669
}
}
logistic-regression:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
8415,
4085
],
[
4190,
8310
]
],
"classification_report": {
"0": {
"precision": 0.6675922253074177,
"recall": 0.6732,
"f1-score": 0.6703843855805617,
"support": 12500
},
"1": {
"precision": 0.6704316256555063,
"recall": 0.6648,
"f1-score": 0.6676039365334404,
"support": 12500
},
"accuracy": 0.669,
"macro avg": {
"precision": 0.6690119254814619,
"recall": 0.669,
"f1-score": 0.668994161057001,
"support": 25000
},
"weighted avg": {
"precision": 0.669011925481462,
"recall": 0.669,
"f1-score": 0.668994161057001,
"support": 25000
}
},
"accuracy_score": 0.669
}
}
logistic-regression-hot-encoding:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
2408,
735
],
[
584,
2523
]
],
"classification_report": {
"0": {
"precision": 0.8048128342245989,
"recall": 0.7661469933184856,
"f1-score": 0.7850040749796251,
"support": 3143
},
"1": {
"precision": 0.7744014732965009,
"recall": 0.8120373350498874,
"f1-score": 0.7927729772191673,
"support": 3107
},
"accuracy": 0.78896,
"macro avg": {
"precision": 0.7896071537605499,
"recall": 0.7890921641841865,
"f1-score": 0.7888885260993962,
"support": 6250
},
"weighted avg": {
"precision": 0.7896947384800229,
"recall": 0.78896,
"f1-score": 0.7888661516609463,
"support": 6250
}
},
"accuracy_score": 0.78896
}
}
random-forest-stemming:
model/metrics.json:
{
"rf_metrics": {
"confusion_matrix": [
[
8322,
4178
],
[
5238,
7262
]
],
"classification_report": {
"0": {
"precision": 0.613716814159292,
"recall": 0.66576,
"f1-score": 0.6386799693016116,
"support": 12500
},
"1": {
"precision": 0.6347902097902098,
"recall": 0.58096,
"f1-score": 0.6066833751044277,
"support": 12500
},
"accuracy": 0.62336,
"macro avg": {
"precision": 0.624253511974751,
"recall": 0.62336,
"f1-score": 0.6226816722030197,
"support": 25000
},
"weighted avg": {
"precision": 0.624253511974751,
"recall": 0.62336,
"f1-score": 0.6226816722030197,
"support": 25000
}
},
"accuracy_score": 0.62336
}
}
svm:
model/metrics.json:
{
"svc_metrics": {
"confusion_matrix": [
[
4539,
7961
],
[
1203,
11297
]
],
"classification_report": {
"0": {
"precision": 0.790491118077325,
"recall": 0.36312,
"f1-score": 0.4976428023243065,
"support": 12500
},
"1": {
"precision": 0.5866133554886281,
"recall": 0.90376,
"f1-score": 0.7114427860696517,
"support": 12500
},
"accuracy": 0.63344,
"macro avg": {
"precision": 0.6885522367829766,
"recall": 0.63344,
"f1-score": 0.6045427941969791,
"support": 25000
},
"weighted avg": {
"precision": 0.6885522367829765,
"recall": 0.63344,
"f1-score": 0.6045427941969792,
"support": 25000
}
},
"accuracy_score": 0.63344
}
}
unigrams-and-bigrams:
model/metrics.json:
{
"lr_metrics": {
"confusion_matrix": [
[
7466,
5034
],
[
6388,
6112
]
],
"classification_report": {
"0": {
"precision": 0.5389057311967663,
"recall": 0.59728,
"f1-score": 0.5665933065189345,
"support": 12500
},
"1": {
"precision": 0.5483581553920689,
"recall": 0.48896,
"f1-score": 0.5169584707772985,
"support": 12500
},
"accuracy": 0.54312,
"macro avg": {
"precision": 0.5436319432944177,
"recall": 0.54312,
"f1-score": 0.5417758886481165,
"support": 25000
},
"weighted avg": {
"precision": 0.5436319432944176,
"recall": 0.54312,
"f1-score": 0.5417758886481165,
"support": 25000
}
},
"accuracy_score": 0.54312
}
}