"gaussian_nb": "GNB",
"random_forest": "RF",
"gradient_boosting": "GB",
"neural_network": "NN",
"knn": "kNN",
"Svm": "SVM",
"random": "Random",
"default": "Default"
}
db = DBHelper()
metadata = pd.DataFrame(db.get_all_metadata(), columns = db.metadata_columns()).drop("id", axis = 1)
models = pd.DataFrame(db.get_all_models(), columns = db.models_columns()).drop("id", axis = 1)
combinations = pd.DataFrame(db.get_all_combinations(), columns = db.combinations_columns())
preperformance = pd.DataFrame(db.get_all_preperformance(), columns = db.preperformance_columns()).drop("id", axis = 1)
# Not null preperformance
preperformance = preperformance[~preperformance.isnull().any(axis = 1)]
preperformance = pd.merge(preperformance, combinations, left_on = "combination_id", right_on = "id").drop(["combination_id", "id", "num_preprocesses"], axis = 1)
models = models.rename(columns = {"model": "classifier"})
models["preprocesses"] = "None"
scores = pd.concat([models, preperformance], sort = False)
scores = scores[scores.classifier != "neural_network"]
models = models[models.classifier != "neural_network"]
metadata_means = {feature: np.mean(metadata[feature]) for feature in metadata.columns if feature != "name"}
metadata.fillna(value = metadata_means, inplace = True)
data = pd.merge(metadata, scores, on = "name")
]
db = DBHelper()
if not os.path.exists("analysis/plots"):
os.makedirs("analysis/plots")
if not os.path.exists("analysis/plots/base_analysis"):
os.makedirs("analysis/plots/base_analysis")
metadata = pd.DataFrame(db.get_all_metadata(),
columns=db.metadata_columns()).drop("id", axis=1)
models = pd.DataFrame(db.get_all_models(),
columns=db.models_columns()).drop("id", axis=1)
combinations = pd.DataFrame(db.get_all_combinations(),
columns=db.combinations_columns())
preperformance = pd.DataFrame(db.get_all_preperformance(),
columns=db.preperformance_columns()).drop("id",
axis=1)
# Not null preperformance
preperformance = preperformance[~preperformance.isnull().any(axis=1)]
preperformance = pd.merge(preperformance,
combinations,
left_on="combination_id",
right_on="id").drop(
["combination_id", "id", "num_preprocesses"],
axis=1)
models = models.rename(columns={"model": "classifier"})
models["preprocesses"] = "None"
scores = pd.concat([models, preperformance], sort=False)
scores = scores[scores.preprocesses.isin(constants.PRE_PROCESSES + ["None"])