| Name | blitzml JSON |
| Version |
0.20.0
JSON |
| download |
| home_page | |
| Summary | A low-code library for machine learning pipelines |
| upload_time | 2023-08-25 12:11:05 |
| maintainer | |
| docs_url | None |
| author | AI Team |
| requires_python | >=3.8 |
| license | |
| keywords |
ml
machine learning
classification
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
<div align="center">
<img src="auxiliary/docs/logo.png" alt="BlitzML" width="400"/>
### **Automate machine learning pipelines rapidly**
<div align="left">
- [Install BlitzML](#install-blitzml)
- [Classification](#classification)
- [Regression](#regression)
- [Time-Series](#time-series)
- [Clustering](#clustering)
# Install BlitzML
```bash
pip install blitzml
```
# Classification
```python
from blitzml.tabular import Classification
import pandas as pd
# prepare your dataframes
train_df = pd.read_csv("auxiliary/datasets/banknote/train.csv")
test_df = pd.read_csv("auxiliary/datasets/banknote/test.csv")
# create the pipeline
auto = Classification(train_df, test_df, algorithm = 'RF', n_estimators = 50)
# perform the entire process
auto.run()
# We can get their values using:
pred_df = auto.pred_df
metrics_dict = auto.metrics_dict
print(pred_df.head())
print(metrics_dict)
```
## Available Classifiers
- Random Forest 'RF'
- LinearDiscriminantAnalysis 'LDA'
- Support Vector Classifier 'SVC'
- KNeighborsClassifier 'KNN'
- GaussianNB 'GNB'
- LogisticRegression 'LR'
- AdaBoostClassifier 'AB'
- GradientBoostingClassifier 'GB'
- DecisionTreeClassifier 'DT'
- MLPClassifier 'MLP'
## **Parameters**
**classifier**
options: {'RF','LDA','SVC','KNN','GNB','LR','AB','GB','DT','MLP', 'auto', 'custom'}, default = 'RF'
`auto: selects the best scoring classifier based on f1-score`
`custom: enables providing a custom classifier through *file_path* and *class_name* parameters`
**file_path**
when using 'custom' classifier, pass the path of the file containing the custom class, default = 'none'
**class_name**
when using 'custom' classifier, pass the class name through this parameter, default = 'none'
**feature_selection**
options: {'correlation', 'importance', 'none'}, default = 'none'
`correlation: use feature columns with the highest correlation with the target`
`importance: use feature columns that are important for the model to predict the target`
`none: use all feature columns`
**validation_percentage**
value determining the validation split percentage (value from 0 to 1), default = 0.1
**average_type**
when performing multiclass classification, provide the average type for the resulting metrics, default = 'macro'
**cross_validation_k_folds**
number of k-folds for cross validation, if 1 then no cv will be performed, default = 1
****kwargs**
optional parameters for the chosen classifier. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html)
## **Attributes**
**train_df**
the preprocessed train dataset (after running `Classification.preprocess()`)
**test_df**
the preprocessed test dataset (after running `Classification.preprocess()`)
**model**
the trained model (after running `Classification.train_the_model()`)
**pred_df**
the prediction dataframe (test_df + predicted target) (after running `Classification.gen_pred_df(Classification.test_df)`)
**metrics_dict**
the validation metrics (after running `Classification.gen_metrics_dict()`)
{
"accuracy": acc,
"f1": f1,
"precision": pre,
"recall": recall,
"hamming_loss": h_loss,
"cross_validation_score":cv_score, `returns None if cross_validation_k_folds==1`
}
## **Methods**
**run()**
a shortcut that runs the entire process:
- preprocessing
- model training
- prediction
- model evaluation
**accuracy_history()**
accuracy scores when varying the sampling size of the train_df (after running `Classification.train_the_model()`).
*returns:*
{
'x':train_df_sample_sizes,
'y1':train_scores_mean,
'y2':test_scores_mean,
'title':title
}
**plot()**
plotting line chart visualizes accuracy history
# Regression
```python
from blitzml.tabular import Regression
import pandas as pd
# prepare your dataframes
train_df = pd.read_csv("auxiliary/datasets/house prices/train.csv")
test_df = pd.read_csv("auxiliary/datasets/house prices/test.csv")
# create the pipeline
auto = Regression(train_df, test_df, algorithm = 'RF')
# perform the entire process
auto.run()
# We can get their values using:
pred_df = auto.pred_df
metrics_dict = auto.metrics_dict
print(pred_df.head())
print(metrics_dict)
```
## Available Regressors
- Random Forest 'RF'
- Support Vector Regressor 'SVR'
- KNeighborsRegressor 'KNN'
- Lasso Regressor 'LSS'
- LinearRegression 'LR'
- Ridge Regressor 'RDG'
- GaussianProcessRegressor 'GPR'
- GradientBoostingRegressor 'GB'
- DecisionTreeRegressor 'DT'
- MLPRegressor 'MLP'
## **Parameters**
**regressor**
options: {'RF','SVR','KNN','LSS','LR','RDG','GPR','GB','DT','MLP', 'auto', 'custom'}, default = 'RF'
`auto: selects the best scoring regressor based on r2 score`
`custom: enables providing a custom regressor through *file_path* and *class_name* parameters`
**file_path**
when using 'custom' regressor, pass the path of the file containing the custom class, default = 'none'
**class_name**
when using 'custom' regressor, pass the class name through this parameter, default = 'none'
**feature_selection**
options: {'correlation', 'importance', 'none'}, default = 'none'
`correlation: use feature columns with the highest correlation with the target`
`importance: use feature columns that are important for the model to predict the target`
`none: use all feature columns`
**validation_percentage**
value determining the validation split percentage (value from 0 to 1), default = 0.1
**cross_validation_k_folds**
number of k-folds for cross validation, if 1 then no cv will be performed, default = 1
****kwargs**
optional parameters for the chosen regressor. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html)
## **Attributes**
**train_df**
the preprocessed train dataset (after running `Regression.preprocess()`)
**test_df**
the preprocessed test dataset (after running `Regression.preprocess()`)
**model**
the trained model (after running `Regression.train_the_model()`)
**pred_df**
the prediction dataframe (test_df + predicted target) (after running `Regression.gen_pred_df(Regression.test_df)`)
**metrics_dict**
the validation metrics (after running `Regression.gen_metrics_dict()`)
{
"r2_score": r2,
"mean_squared_error": mse,
"root_mean_squared_error": rmse,
"mean_absolute_error" : mae,
"cross_validation_score":cv_score, `returns None if cross_validation_k_folds==1`
}
## **Methods**
**run()**
a shortcut that runs the entire process:
- preprocessing
- model training
- prediction
- model evaluation
**plot()**
plotting line chart visualizes RMSE history
**RMSE_history()**
RMSE scores when varying the sampling size of the train_df (after running `Regression.train_the_model()`).
*returns:*
{
'x':train_df_sample_sizes,
'y1':train_scores_mean,
'y2':test_scores_mean,
'title':title
}
# Time-series
time series is a particular problem of Regression, but time series have some additional functions:
- stationary test (IsStationary()).
- convert to stationary.
- reverse predicted.
and the dataset must have a DateTime column, even if the DataType of this column is Object.
```python
from blitzml.tabular import TimeSeries
import pandas as pd
# prepare your dataframes
train_df = pd.read_csv("train_dataset.csv")
test_df = pd.read_csv("test_dataset.csv")
# create the pipeline
auto = TimeSeries(train_df, test_df, algorithm = 'RF')
# Perform the entire process:
auto.run()
# We can get their values using:
pred_df = auto.pred_df
metrics_dict = auto.metrics_dict
print(pred_df.head())
print(metrics_dict)
```
# Clustering
```python
from blitzml.unsupervised import Clustering
import pandas as pd
# prepare your dataframe
train_df = pd.read_csv("auxiliary/datasets/customer personality/train.csv")
# create the pipeline
auto = Clustering(train_df, clustering_algorithm = 'KM')
# first perform data preprocessing
auto.preprocess()
# second train the model
auto.train_the_model()
# After training the model we can generate:
auto.gen_pred_df()
auto.gen_metrics_dict()
# We can get their values using:
print(auto.pred_df.head())
print(auto.metrics_dict)
```
## Available Clustering Algorithms
- K-Means 'KM'
- Affinity Propagation 'AP'
- Agglomerative Clustering 'AC'
- Mean Shift 'MS'
- Spectral Clustering 'SC'
- Birch 'Birch'
- Bisecting K-Means 'BKM'
- OPTICS 'OPTICS'
- DBSCAN 'DBSCAN'
## **Parameters**
**clustering_algorithm**
options: {"KM", "AP", "AC", "MS", "SC", "Birch", "BKM", "OPTICS", "DBSCAN", 'auto', 'custom'}, default = 'KM'
`auto: selects the best scoring clustering algorithm based on silhouette score`
`custom: enables providing a custom clustering algorithm through *file_path* and *class_name* parameters`
**file_path**
when using 'custom' clustering_algorithm, pass the path of the file containing the custom class, default = 'none'
**class_name**
when using 'custom' clustering_algorithm, pass the class name through this parameter, default = 'none'
**feature_selection**
options: {'importance', 'none'}, default = 'none'
`importance: use feature columns that are important for the model to predict the target`
`none: use all feature columns`
****kwargs**
optional parameters for the chosen clustering_algorithm. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html)
## **Attributes**
**train_df**
the preprocessed train dataset (after running `Clustering.preprocess()`)
**model**
the trained model (after running `Clustering.train_the_model()`)
**pred_df**
the prediction dataframe (test_df + predicted target) (after running `Clustering.gen_pred_df()`)
**metrics_dict**
the validation metrics (after running `Clustering.gen_metrics_dict()`)
{
"silhouette_score": sil_score,
"calinski_harabasz_score": cal_har_score,
"davies_bouldin_score": dav_boul_score,
"n_clusters" : n
}
## **Methods**
**preprocess()**
perform preprocessing on train_df
**train_the_model()**
train the chosen clustering algorithm on the train_df
**clustering_visualization()**
2-d visualization of the data points with its corresponding labels (after doing dimensionality reduction using Principal Componenet Analysis).
*returns:*
{
'principal_component_1':pc1,
'principal_component_2':pc2,
'cluster_labels':labels,
'title':title
}
**gen_pred_df()**
generates the prediction dataframe and assigns it to the `pred_df` attribute
**gen_metrics_dict()**
generates the clustering metrics and assigns it to the `metrics_dict`
**run()**
a shortcut that runs the following methods:
preprocess()
train_the_model()
gen_pred_df()
gen_metrics_dict()
## Development
- Clone the repo
- run `pip install virtualenv`
- run `python -m virtualenv venv`
- run `. ./venv/bin/activate` on UNIX based systems or `. ./venv/Scripts/activate.ps1` if on windows
- run `pip install -r requirements.txt`
- run `pre-commit install`
Raw data
{
"_id": null,
"home_page": "",
"name": "blitzml",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.8",
"maintainer_email": "",
"keywords": "ml,machine learning,classification",
"author": "AI Team",
"author_email": "AI Team <CV.Team.CSE.2023@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/76/b7/b5a5255509c646dcef4fd75ebeb182b76b7618f9bc67d816768871c5ecc0/blitzml-0.20.0.tar.gz",
"platform": null,
"description": "\r\n<div align=\"center\">\r\n<img src=\"auxiliary/docs/logo.png\" alt=\"BlitzML\" width=\"400\"/>\r\n\r\n### **Automate machine learning pipelines rapidly**\r\n\r\n\r\n<div align=\"left\">\r\n\r\n- [Install BlitzML](#install-blitzml)\r\n- [Classification](#classification)\r\n- [Regression](#regression)\r\n- [Time-Series](#time-series)\r\n- [Clustering](#clustering)\r\n\r\n\r\n\r\n# Install BlitzML \r\n\r\n\r\n```bash\r\npip install blitzml\r\n```\r\n\r\n\r\n# Classification\r\n\r\n```python\r\nfrom blitzml.tabular import Classification\r\nimport pandas as pd\r\n\r\n# prepare your dataframes\r\ntrain_df = pd.read_csv(\"auxiliary/datasets/banknote/train.csv\")\r\ntest_df = pd.read_csv(\"auxiliary/datasets/banknote/test.csv\")\r\n\r\n# create the pipeline\r\nauto = Classification(train_df, test_df, algorithm = 'RF', n_estimators = 50)\r\n\r\n# perform the entire process\r\nauto.run()\r\n\r\n# We can get their values using:\r\npred_df = auto.pred_df\r\nmetrics_dict = auto.metrics_dict\r\n\r\nprint(pred_df.head())\r\nprint(metrics_dict)\r\n```\r\n\r\n\r\n## Available Classifiers\r\n\r\n- Random Forest 'RF' \r\n- LinearDiscriminantAnalysis 'LDA' \r\n- Support Vector Classifier 'SVC' \r\n- KNeighborsClassifier 'KNN' \r\n- GaussianNB 'GNB' \r\n- LogisticRegression 'LR'\r\n- AdaBoostClassifier 'AB'\r\n- GradientBoostingClassifier 'GB'\r\n- DecisionTreeClassifier 'DT'\r\n- MLPClassifier 'MLP'\r\n\r\n\r\n## **Parameters**\r\n**classifier** \r\noptions: {'RF','LDA','SVC','KNN','GNB','LR','AB','GB','DT','MLP', 'auto', 'custom'}, default = 'RF' \r\n`auto: selects the best scoring classifier based on f1-score` \r\n`custom: enables providing a custom classifier through *file_path* and *class_name* parameters` \r\n**file_path** \r\nwhen using 'custom' classifier, pass the path of the file containing the custom class, default = 'none' \r\n**class_name** \r\nwhen using 'custom' classifier, pass the class name through this parameter, default = 'none' \r\n**feature_selection** \r\noptions: {'correlation', 'importance', 'none'}, default = 'none' \r\n`correlation: use feature columns with the highest correlation with the target` \r\n`importance: use feature columns that are important for the model to predict the target` \r\n`none: use all feature columns` \r\n**validation_percentage** \r\nvalue determining the validation split percentage (value from 0 to 1), default = 0.1 \r\n**average_type** \r\nwhen performing multiclass classification, provide the average type for the resulting metrics, default = 'macro' \r\n**cross_validation_k_folds** \r\nnumber of k-folds for cross validation, if 1 then no cv will be performed, default = 1 \r\n****kwargs** \r\noptional parameters for the chosen classifier. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html) \r\n## **Attributes** \r\n**train_df** \r\nthe preprocessed train dataset (after running `Classification.preprocess()`) \r\n**test_df** \r\nthe preprocessed test dataset (after running `Classification.preprocess()`) \r\n**model** \r\nthe trained model (after running `Classification.train_the_model()`) \r\n**pred_df** \r\nthe prediction dataframe (test_df + predicted target) (after running `Classification.gen_pred_df(Classification.test_df)`) \r\n**metrics_dict** \r\nthe validation metrics (after running `Classification.gen_metrics_dict()`) \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0\"accuracy\": acc, \r\n\u00a0\u00a0\u00a0\u00a0\"f1\": f1, \r\n\u00a0\u00a0\u00a0\u00a0\"precision\": pre, \r\n\u00a0\u00a0\u00a0\u00a0\"recall\": recall, \r\n\u00a0\u00a0\u00a0\u00a0\"hamming_loss\": h_loss, \r\n\u00a0\u00a0\u00a0\u00a0\"cross_validation_score\":cv_score, `returns None if cross_validation_k_folds==1` \r\n} \r\n## **Methods** \r\n**run()** \r\na shortcut that runs the entire process: \r\n- preprocessing\r\n- model training \r\n- prediction \r\n- model evaluation \r\n\r\n**accuracy_history()** \r\naccuracy scores when varying the sampling size of the train_df (after running `Classification.train_the_model()`). \r\n*returns:* \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0'x':train_df_sample_sizes, \r\n\u00a0\u00a0\u00a0\u00a0'y1':train_scores_mean, \r\n\u00a0\u00a0\u00a0\u00a0'y2':test_scores_mean, \r\n\u00a0\u00a0\u00a0\u00a0'title':title \r\n} \r\n**plot()**\r\n\r\nplotting line chart visualizes accuracy history\r\n\r\n\r\n# Regression \r\n\r\n```python\r\nfrom blitzml.tabular import Regression\r\nimport pandas as pd\r\n\r\n# prepare your dataframes\r\ntrain_df = pd.read_csv(\"auxiliary/datasets/house prices/train.csv\")\r\ntest_df = pd.read_csv(\"auxiliary/datasets/house prices/test.csv\")\r\n\r\n# create the pipeline\r\nauto = Regression(train_df, test_df, algorithm = 'RF')\r\n\r\n# perform the entire process\r\nauto.run()\r\n\r\n# We can get their values using:\r\npred_df = auto.pred_df\r\nmetrics_dict = auto.metrics_dict\r\n\r\nprint(pred_df.head())\r\nprint(metrics_dict)\r\n```\r\n\r\n\r\n## Available Regressors\r\n\r\n- Random Forest 'RF'\r\n- Support Vector Regressor 'SVR'\r\n- KNeighborsRegressor 'KNN'\r\n- Lasso Regressor 'LSS'\r\n- LinearRegression 'LR'\r\n- Ridge Regressor 'RDG'\r\n- GaussianProcessRegressor 'GPR'\r\n- GradientBoostingRegressor 'GB'\r\n- DecisionTreeRegressor 'DT'\r\n- MLPRegressor 'MLP'\r\n\r\n## **Parameters**\r\n**regressor** \r\noptions: {'RF','SVR','KNN','LSS','LR','RDG','GPR','GB','DT','MLP', 'auto', 'custom'}, default = 'RF' \r\n`auto: selects the best scoring regressor based on r2 score` \r\n`custom: enables providing a custom regressor through *file_path* and *class_name* parameters` \r\n**file_path** \r\nwhen using 'custom' regressor, pass the path of the file containing the custom class, default = 'none' \r\n**class_name** \r\nwhen using 'custom' regressor, pass the class name through this parameter, default = 'none' \r\n**feature_selection** \r\noptions: {'correlation', 'importance', 'none'}, default = 'none' \r\n`correlation: use feature columns with the highest correlation with the target` \r\n`importance: use feature columns that are important for the model to predict the target` \r\n`none: use all feature columns` \r\n**validation_percentage** \r\nvalue determining the validation split percentage (value from 0 to 1), default = 0.1 \r\n**cross_validation_k_folds** \r\nnumber of k-folds for cross validation, if 1 then no cv will be performed, default = 1 \r\n****kwargs** \r\noptional parameters for the chosen regressor. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html) \r\n## **Attributes** \r\n**train_df** \r\nthe preprocessed train dataset (after running `Regression.preprocess()`) \r\n**test_df** \r\nthe preprocessed test dataset (after running `Regression.preprocess()`) \r\n**model** \r\nthe trained model (after running `Regression.train_the_model()`) \r\n**pred_df** \r\nthe prediction dataframe (test_df + predicted target) (after running `Regression.gen_pred_df(Regression.test_df)`) \r\n**metrics_dict** \r\nthe validation metrics (after running `Regression.gen_metrics_dict()`) \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0\"r2_score\": r2, \r\n\u00a0\u00a0\u00a0\u00a0\"mean_squared_error\": mse, \r\n\u00a0\u00a0\u00a0\u00a0\"root_mean_squared_error\": rmse, \r\n\u00a0\u00a0\u00a0\u00a0\"mean_absolute_error\" : mae, \r\n\u00a0\u00a0\u00a0\u00a0\"cross_validation_score\":cv_score, `returns None if cross_validation_k_folds==1` \r\n} \r\n## **Methods** \r\n\r\n\r\n**run()** \r\na shortcut that runs the entire process: \r\n- preprocessing\r\n- model training \r\n- prediction \r\n- model evaluation \r\n\r\n**plot()**\r\n\r\nplotting line chart visualizes RMSE history\r\n\r\n**RMSE_history()** \r\nRMSE scores when varying the sampling size of the train_df (after running `Regression.train_the_model()`). \r\n*returns:* \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0'x':train_df_sample_sizes, \r\n\u00a0\u00a0\u00a0\u00a0'y1':train_scores_mean, \r\n\u00a0\u00a0\u00a0\u00a0'y2':test_scores_mean, \r\n\u00a0\u00a0\u00a0\u00a0'title':title \r\n} \r\n# Time-series\r\ntime series is a particular problem of Regression, but time series have some additional functions:\r\n- stationary test (IsStationary()). \r\n- convert to stationary.\r\n- reverse predicted.\r\n\r\nand the dataset must have a DateTime column, even if the DataType of this column is Object.\r\n```python\r\nfrom blitzml.tabular import TimeSeries \r\nimport pandas as pd\r\n\r\n# prepare your dataframes\r\ntrain_df = pd.read_csv(\"train_dataset.csv\")\r\ntest_df = pd.read_csv(\"test_dataset.csv\")\r\n\r\n# create the pipeline\r\nauto = TimeSeries(train_df, test_df, algorithm = 'RF')\r\n\r\n# Perform the entire process:\r\nauto.run()\r\n\r\n# We can get their values using:\r\npred_df = auto.pred_df\r\nmetrics_dict = auto.metrics_dict\r\n\r\nprint(pred_df.head())\r\nprint(metrics_dict)\r\n```\r\n# Clustering \r\n\r\n```python\r\nfrom blitzml.unsupervised import Clustering\r\nimport pandas as pd\r\n\r\n# prepare your dataframe\r\ntrain_df = pd.read_csv(\"auxiliary/datasets/customer personality/train.csv\")\r\n\r\n# create the pipeline\r\nauto = Clustering(train_df, clustering_algorithm = 'KM')\r\n\r\n# first perform data preprocessing\r\nauto.preprocess()\r\n# second train the model\r\nauto.train_the_model()\r\n\r\n# After training the model we can generate:\r\nauto.gen_pred_df()\r\nauto.gen_metrics_dict()\r\n\r\n# We can get their values using:\r\nprint(auto.pred_df.head())\r\nprint(auto.metrics_dict)\r\n```\r\n\r\n\r\n## Available Clustering Algorithms \r\n\r\n- K-Means 'KM' \r\n- Affinity Propagation 'AP' \r\n- Agglomerative Clustering 'AC' \r\n- Mean Shift 'MS' \r\n- Spectral Clustering 'SC' \r\n- Birch 'Birch' \r\n- Bisecting K-Means 'BKM' \r\n- OPTICS 'OPTICS' \r\n- DBSCAN 'DBSCAN' \r\n\r\n## **Parameters** \r\n**clustering_algorithm** \r\noptions: {\"KM\", \"AP\", \"AC\", \"MS\", \"SC\", \"Birch\", \"BKM\", \"OPTICS\", \"DBSCAN\", 'auto', 'custom'}, default = 'KM' \r\n`auto: selects the best scoring clustering algorithm based on silhouette score` \r\n`custom: enables providing a custom clustering algorithm through *file_path* and *class_name* parameters` \r\n**file_path** \r\nwhen using 'custom' clustering_algorithm, pass the path of the file containing the custom class, default = 'none' \r\n**class_name**\r\nwhen using 'custom' clustering_algorithm, pass the class name through this parameter, default = 'none' \r\n**feature_selection** \r\noptions: {'importance', 'none'}, default = 'none' \r\n`importance: use feature columns that are important for the model to predict the target` \r\n`none: use all feature columns` \r\n****kwargs** \r\noptional parameters for the chosen clustering_algorithm. you can find available parameters in the [sklearn docs](https://scikit-learn.org/stable/user_guide.html) \r\n## **Attributes** \r\n**train_df** \r\nthe preprocessed train dataset (after running `Clustering.preprocess()`) \r\n**model** \r\nthe trained model (after running `Clustering.train_the_model()`) \r\n**pred_df** \r\nthe prediction dataframe (test_df + predicted target) (after running `Clustering.gen_pred_df()`) \r\n**metrics_dict** \r\nthe validation metrics (after running `Clustering.gen_metrics_dict()`) \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0\"silhouette_score\": sil_score, \r\n\u00a0\u00a0\u00a0\u00a0\"calinski_harabasz_score\": cal_har_score, \r\n\u00a0\u00a0\u00a0\u00a0\"davies_bouldin_score\": dav_boul_score, \r\n\u00a0\u00a0\u00a0\u00a0\"n_clusters\" : n \r\n} \r\n## **Methods** \r\n**preprocess()** \r\nperform preprocessing on train_df \r\n**train_the_model()** \r\ntrain the chosen clustering algorithm on the train_df \r\n**clustering_visualization()** \r\n2-d visualization of the data points with its corresponding labels (after doing dimensionality reduction using Principal Componenet Analysis). \r\n*returns:* \r\n{ \r\n\u00a0\u00a0\u00a0\u00a0'principal_component_1':pc1, \r\n\u00a0\u00a0\u00a0\u00a0'principal_component_2':pc2, \r\n\u00a0\u00a0\u00a0\u00a0'cluster_labels':labels, \r\n\u00a0\u00a0\u00a0\u00a0'title':title \r\n} \r\n**gen_pred_df()** \r\ngenerates the prediction dataframe and assigns it to the `pred_df` attribute \r\n**gen_metrics_dict()** \r\ngenerates the clustering metrics and assigns it to the `metrics_dict` \r\n**run()** \r\na shortcut that runs the following methods: \r\npreprocess() \r\ntrain_the_model() \r\ngen_pred_df() \r\ngen_metrics_dict() \r\n## Development \r\n\r\n- Clone the repo \r\n- run `pip install virtualenv`\r\n- run `python -m virtualenv venv`\r\n- run `. ./venv/bin/activate` on UNIX based systems or `. ./venv/Scripts/activate.ps1` if on windows\r\n- run `pip install -r requirements.txt`\r\n- run `pre-commit install`\r\n",
"bugtrack_url": null,
"license": "",
"summary": "A low-code library for machine learning pipelines",
"version": "0.20.0",
"project_urls": {
"Homepage": "https://github.com/AhmedMohamed25/blitzml"
},
"split_keywords": [
"ml",
"machine learning",
"classification"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "dd85b8b07124cebf6573f3ea11d0ca4edaf2fd5fd18f7be81c5c558f136cf33b",
"md5": "5a1806498d66d64fbd9f75c03c507884",
"sha256": "779e6726bc27382a72d0c07693c46fcea8b5875ba5f54c9a11b4d348623b9ee0"
},
"downloads": -1,
"filename": "blitzml-0.20.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "5a1806498d66d64fbd9f75c03c507884",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.8",
"size": 19336,
"upload_time": "2023-08-25T12:11:03",
"upload_time_iso_8601": "2023-08-25T12:11:03.040231Z",
"url": "https://files.pythonhosted.org/packages/dd/85/b8b07124cebf6573f3ea11d0ca4edaf2fd5fd18f7be81c5c558f136cf33b/blitzml-0.20.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "76b7b5a5255509c646dcef4fd75ebeb182b76b7618f9bc67d816768871c5ecc0",
"md5": "7d4ecd29b64ae76f03bd8af9ee0242a7",
"sha256": "ef023f48b1ef50ccfbb6601db28074f638a6dd5abce2f45960fee35515f87fde"
},
"downloads": -1,
"filename": "blitzml-0.20.0.tar.gz",
"has_sig": false,
"md5_digest": "7d4ecd29b64ae76f03bd8af9ee0242a7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.8",
"size": 16654,
"upload_time": "2023-08-25T12:11:05",
"upload_time_iso_8601": "2023-08-25T12:11:05.119291Z",
"url": "https://files.pythonhosted.org/packages/76/b7/b5a5255509c646dcef4fd75ebeb182b76b7618f9bc67d816768871c5ecc0/blitzml-0.20.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-08-25 12:11:05",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "AhmedMohamed25",
"github_project": "blitzml",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"requirements": [],
"lcname": "blitzml"
}
