How to Build a ML Model in 1 Minute using ChatGPT

Nitika Sharma 06 May, 2024
6 min read


Machine learning (ML) has become a game-changer across industries, but its complexity can be intimidating. This article explores how to use ChatGPT to build machine learning models. We’ll look into how ChatGPT can assist in various stages of model creation, from data preparation to training and evaluation, all through an intuitive conversational interface.

Why use ChatGPT for Building Machine Learning Models?

Unlike traditional tools that require extensive coding expertise, ChatGPT utilizes a user-friendly conversational interface. This allows you to interact with ChatGPT naturally, asking questions and receiving guidance on various stages of model creation. From brainstorming initial problem definitions to cleaning and preparing data, ChatGPT can be your AI partner.

It can assist with tasks like feature engineering, where it helps identify relevant features from raw data to improve model performance. It can even offer insights into model evaluation, suggesting metrics and interpreting results to guide your decision-making.

By leveraging ChatGPT’s conversational power, you can potentially save valuable time and resources throughout the ML development process. Whether you’re a seasoned data scientist or just starting your ML journey, ChatGPT can be a valuable asset in building robust and effective models.

Also Read: Top 10 Machine Learning Algorithms to Use in 2024

Steps Involved in Building ML Model using ChatGPT

While ChatGPT can’t single-handedly build a machine learning model, it can act as a powerful conversational guide throughout the process. Here’s a breakdown of the typical steps involved in building a machine learning model, along with how ChatGPT can assist:

Problem Definition

Describe your objective to ChatGPT. ChatGPT can help brainstorm potential applications of machine learning and refine your problem statement.

Data Collection

Explain the type of data you need to ChatGPT. ChatGPT can suggest potential data sources and help identify relevant data formats.

Data Cleaning and Preprocessing

Describe any data quality issues you encounter. ChatGPT can guide you through data cleaning techniques like handling missing values or outliers.

Data Exploration and Feature Engineering

Upload your data (if possible) and ask ChatGPT to analyze it. ChatGPT can help identify patterns, suggest potential features for your model, and perform basic feature engineering tasks.

Model Selection and Training

Explain the type of problem you’re trying to solve (classification, regression, etc.). Based on your problem and data, ChatGPT can recommend suitable machine learning algorithms and guide you through basic model training steps.

Model Evaluation

Provide ChatGPT with your model’s evaluation metrics (accuracy, precision, recall, etc.). ChatGPT can help you interpret the results and suggest strategies for improving model performance.

Model Deployment and Monitoring

While ChatGPT can’t directly deploy your model, it can help you understand deployment considerations and suggest tools or platforms. Remember: Throughout these steps, ChatGPT acts as your conversational co-pilot. It can’t write code or train models on its own, but it can provide valuable guidance, suggest resources, and streamline the process by facilitating clear communication of your goals and challenges.

Also Read: Machine Learning Basics for a Newbie

Let’s Build a Machine Learning Model with ChatGPT

In this section, we will look at a basic example of building a machine learning model with ChatGPT. We will follow all the steps mentioned above.

Data Collection and Cleaning


Load the Cancer Dataset from statsmodel, and build a complete Machine Learning Model using this dataset for classification. First, define the problem statement for me and secondly, perform data cleaning and preprocessing task like encoding, checking missing values and handling them also outlier treatment.

Code Generated by ChatGPT

# Importing necessary libraries
import numpy as np
import pandas as pd
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, classification_report, confusion_matrix

# Load the breast cancer dataset
data = load_breast_cancer()

# Converting to DataFrame
df = pd.DataFrame(, columns=data.feature_names)
df['target'] =

# Checking for missing values
print("Missing values:\n", df.isnull().sum())

# No missing values found

# Checking for outliers
print("\nOutliers (if any):\n", df.describe())

# No obvious outliers found, but we'll proceed with outlier treatment if necessary.

# Encoding categorical variables (if any)
# There are no categorical variables to encode as all features are numerical.


# Checking for outliers
# No obvious outliers found, but we'll proceed with outlier treatment if necessary.

# Encoding categorical variables (if any)
# There are no categorical variables to encode as all features are numerical.

Data Exploration and Feature Engineering


Now perform Data Exploration and Feature Engineering this on the same dataset.

Code Generated by ChatGPT

import matplotlib.pyplot as plt
import seaborn as sns

# Data Exploration

# Check the distribution of target variable
plt.figure(figsize=(6, 4))
sns.countplot(x='target', data=df)
plt.title('Distribution of Target Variable')

# Check the distribution of features
plt.figure(figsize=(14, 10))
for i, feature in enumerate(df.columns[:-1]):
    plt.subplot(5, 6, i+1)
    sns.histplot(df[feature], kde=True)

# Analyze correlations between features and target variable
plt.figure(figsize=(14, 10))
correlation_matrix = df.corr()
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt=".2f")
plt.title('Correlation Matrix')

# Feature Engineering

# Feature selection - selecting features with high absolute correlation with the target
important_features = correlation_matrix['target'].apply(abs).sort_values(ascending=False).index[1:11]

# Selecting the important features
X_selected = X[important_features]

# Splitting data into training and testing sets
X_train_selected, X_test_selected, _, _ = train_test_split(X_selected, y, test_size=0.2, random_state=42)

# Standardizing selected features
scaler_selected = StandardScaler()
X_train_selected_scaled = scaler_selected.fit_transform(X_train_selected)
X_test_selected_scaled = scaler_selected.transform(X_test_selected)


Count Plot
Sub-plot and Histogram plot
Correlation Matrix

Model Selection and Evaluation


Now perform Model Selection and Training on the scaled dataset created in previous step/ prompt.

from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.model_selection import cross_val_score

# Initialize classifiers
classifiers = {
    "Logistic Regression": LogisticRegression(),
    "Support Vector Machine": SVC(),
    "K-Nearest Neighbors": KNeighborsClassifier(),
    "Decision Tree": DecisionTreeClassifier(),
    "Random Forest": RandomForestClassifier(),
    "Gradient Boosting": GradientBoostingClassifier()

# Training and evaluating models
results = {}
for name, clf in classifiers.items():
    scores = cross_val_score(clf, X_train_selected_scaled, y_train, cv=5, scoring='accuracy')
    results[name] = scores.mean()

# Display mean accuracy of each model
print("\nMean Accuracy of Models:")
for name, score in results.items():
    print(f"{name}: {score:.4f}")

# Select the best performing model
best_model_name = max(results, key=results.get)
best_model = classifiers[best_model_name]

# Train the best model, y_train)

# Predicting on the test set with the best model
y_pred_best = best_model.predict(X_test_selected_scaled)

# Evaluating the best model
accuracy_best = accuracy_score(y_test, y_pred_best)
print("\nBest Model:", best_model_name)
print("\nAccuracy with the best model:", accuracy_best)
print("\nClassification Report with the best model:\n", classification_report(y_test, y_pred_best))
print("\nConfusion Matrix with the best model:\n", confusion_matrix(y_test, y_pred_best))


Model Evaluation

And with this, we have successfuly built a ML model in less than 1 minute.


In conclusion, while ChatGPT has been discussed here, any AI chatbot can serve a similar purpose in building machine learning models. Moreover, AI chatbots can handle various other machine learning tasks beyond model creation.

Have you tried using AI for tasks like data analysis or natural language processing? Let me know in the comment section below!

For more such engaging content, keep following Analytics Vidhya Blogs!

Frequently Asked Questions

Q1. Can ChatGPT create ML models?

A. No, ChatGPT can’t create ML models on its own, but it can guide you through the process, offering assistance at every step.

Q2. Can ChatGPT do machine learning?

A. No, ChatGPT itself doesn’t do machine learning, but it can help with tasks like data preprocessing, model selection, and evaluation in ML projects.

Q3. How to use ChatGPT for machine learning projects?

A. You can use ChatGPT by interacting with it naturally, asking for guidance on problem definition, data cleaning, model selection, evaluation, and deployment.

Q4. How do I create a custom machine learning model?

A. To create a custom ML model, first define your problem, collect and preprocess data, select suitable algorithms, train the model, evaluate its performance, and finally deploy it. ChatGPT can assist you at every stage.

Nitika Sharma 06 May, 2024

Frequently Asked Questions

Lorem ipsum dolor sit amet, consectetur adipiscing elit,

Responses From Readers