AutoML

About the Project

AutoML Builder is a Python project that automatically generates a complete machine learning model from a CSV file. It handles data preprocessing, encoding, model selection, and evaluation with minimal user input.

Features

• Automatic detection of feature types (numerical, categorical, ordinal)
• Handles missing values and outliers
• Selects and trains the best ML model
• Provides evaluation metrics for regression and classification
• Easy integration with custom pipelines

Installation

pip install mlaunch

How to use

1. create a python file and paste this code:

import pandas as pd
import warnings
import os
import subprocess

from mlaunch import AutoML

warnings.filterwarnings("ignore")


path = input("enter the path for the dataset: ")
df = pd.read_csv(path)
for column in df.columns.tolist():
    print(column)
y_column = input("choose the y column in the dataframe: ")
models_names = ["Linear Regression","Logistic Regression","Random Forest Regression","Hist Gradient Boosting Regression","Random Forest classifier","Hist Gradient Boosting classifier","Auto Select"]
for model_name in models_names:
    print(f"{models_names.index(model_name) + 1}- ",model_name)
model_name = models_names[int(input("choose a model by writing the number crossponding with the model you want: "))-1]

folder_path,score = AutoML(path,y_column,model_name)
print("model craeted sucessfully 🥳")
print(f"path: {folder_path}")
print(f"score: {score}")
subprocess.run(["python", os.path.join(folder_path,"ML_model.py")], check=True)

package

Functions

AutoML

this function will preprocess the data and create the model for you

from mlaunch import AutoML
model = AutoML(path,y_column,model_name,type = "pipeline")

parameteres

• path: the path for your csv file
• y_column: the target column in your dataset
• model_name: the name of your model out of these current model:
  • Linear Regression
  • Logistic Regression
  • Random Forest Regression
  • Hist Gradient Boosting Regression
  • Random Forest classifier
  • Hist Gradient Boosting classifier
  • Auto Select : it will select the best model for your data more will be added in the future
• type: how will the output be:
  • pipeline: it will output the model as a pipeline
  • python file: it will export the model to model.pkl file and run a python file to input the data

preprocessing

this function will handle the outliers and encode your data and output it as a ColumnTransformer

from mlaunch import preprocessing
preprocessor = preprocessing(model,df,y_column)
model = Pipeline([
    ("preprocessor",preprocessor)
])

parameteres

• model: put your model here
• df: put the dataframe here
• y_column: the target column in your dataset

dataset_info

this function returns a dictionary of the size, cat_columns and num_columns

from mlaunch import dataset_info
info = dataset_info(df,y_column)

parameteres

• df: put the dataframe here
• y_column: the target column in your dataset

column_statistics

this function returns a bunch of stats for every column in the dataframe

from mlaunch import column_statistics
stats = column_statistics(df,cat_columns,num_columns)

parameteres

• df: put the dataframe here
• cat_columns: the catagorical columns in your dataframe, you can get them easily by using dataset_info(df,y_column)["cat_columns]
• num_columns: the numerical columns in your dataframe, you can get them easily by using dataset_info(df,y_column)["num_columns]