Photo by Author | Chat GPT
. Introduction
When it comes to machine learning, the performance is key. Writing clean, readable, and comprehensive code not only accelerates growth, but also makes your machine learning pipelines easier to understand, distribute, maintain and debug. Interesting, with its natural and expressive combination, a powerful one liner is a great fit to produce the common tasks in just one line of the code.
This tutorial will focus on ten practical one liners who take advantage of the power of libraries Skate And Pandas Help your machine learning workflows to smooth. We will cover everything from data preparation and model training to diagnosis and feature analysis.
Let’s start.
. Establish
Before preparing your code, let’s import the necessary libraries that we will use for example.
import pandas as pd
from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.pipeline import Pipeline
from sklearn.datasets import load_iris
from sklearn.metrics import accuracy_scoreBeyond this route, let’s code … a line at a time.
. 1. Datastate loading
Let’s start with one of the basics. Starting with a project often means loading data. Skate Learn comes with several toy datases that are perfect for models and workflow testing. You can load both features and target variables in the same, clean line.
X, y = load_iris(return_X_y=True)It uses one liner load_iris Function and set return_X_y=True To return the feature matrix directly X And target vector yAvoid the need to analyze an item like a dictionary.
. 2. Divide data into training and testing sets
Another basic step of any machine learning project is to divide your data into various Multiple sets of various use. train_test_split The function is an important place. It can be hanged in one line to create four separate data fames for your training and test sets.
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42, stratify=y)Here, we use test_size=0.3 To allocate and use 30 30 % data of test stratify=y The original dataset is a mirror to ensure the proportion of classes in the train and test sets.
. 3. Creating and training a model
Why use two lines to speed up the model and then train it? You can make the chain fit Direct method of the model conductor for a compact and reading line of code, thus:
model = LogisticRegression(max_iter=1000, random_state=42).fit(X_train, y_train)This single line produces A LogisticRegression The model and immediately trains it on its training data, returning the footed model object.
. 4. K-Fold Cross Verification
Cross verification is divided into a train test far more than your model’s performance. Of skate learns cross_val_score This diagnosis makes it easy to perform in one step.
scores = cross_val_score(LogisticRegression(max_iter=1000, random_state=42), X, y, cv=5)This one -liner initiates a new logistics registration model, divides the data 5 times, divides into trains and reviews the model 5 times (((cv=5), And from each fold, a list of scores.
. 5. Predictions and accuracy of accuracy
After training your model, you would like to assess its performance on the test set. You can do this and get a score of accuracy with the same procedure.
accuracy = model.score(X_test, y_test) .score() The method easily connects the measures of the model, while returning the model accuracy on the test data provided.
. 6. Scaling numeric properties
Feature Skilling is a common pre -processing move, especially for sensitive algorithms on a scale of input properties – including SVM and logistic regulation. You can fit the scaller and change your data simultaneously using this single line of Ajgar:
X_scaled = StandardScaler().fit_transform(X) fit_transform The method is an easy shortcut that learns data from data and applies change at the same time.
. 7. A hot encoding on dotok data
A standard technique to handle a hot encoding category properties. While skate learn is a powerful OneHotEncoder Method powerful, get_dummies The function from pandas allows a true one liner for this task.
df_encoded = pd.get_dummies(pd.DataFrame(X, columns=('f1', 'f2', 'f3', 'f4')), columns=('f1'))This line changes a specific column (f1) In new columns with binary values ​​in pandas data frame (f1, f2, f3, f4), Best for machine learning model.
. 8. to describe Skyk Learn Pipeline
Scaten Learn pipelines make multiple processing stages and one final estimated straight to the one. They prevent data leakage and simplify your workflow. To specify the pipeline is a clean liner, such as the following:
pipeline = Pipeline((('scaler', StandardScaler()), ('svc', SVC())))This produces a pipeline that scales the data by first use StandardScaler And then feeds the result in the support vector rating.
. 9. Tunu the hyperpressors with Grid Search CV
Finding the best hyperpresster for your model can be painful. GridSearchCV Can help make this process automatic. From chains .fit()You can start, explain the search, and it can all run in the same line.
grid_search = GridSearchCV(SVC(), {'C': (0.1, 1, 10), 'kernel': ('linear', 'rbf')}, cv=3).fit(X_train, y_train)This makes a grid looking for an AN SVC Model, tests different values C And kernel3 -fold cross -verification performs (cv=3), And fits training data to find the best combination.
. 10. Extract the import of feature
For trees -oriented models such as random forests, it is very important to understand what features are the most influential and effective model. A listing feature is a classic one -liner to extract and configure imports of a list. Note that this quote first forms a model and then uses one liner to determine the import of features.
# First, train a model
feature_names = ('sepal_length', 'sepal_width', 'petal_length', 'petal_width')
rf_model = RandomForestClassifier(random_state=42).fit(X_train, y_train)
# The one-liner
importances = sorted(zip(feature_names, rf_model.feature_importances_), key=lambda x: x(1), reverse=True)This one -liner connects the name of each feature with its importance, then sets the list in a descending order to showcase the most important features.
. Wrap
These ten one -liners show how a comprehensive syntax of Azar can help you write a more efficient and reading machine learning code. Integrate these shortcuts into your daily workflows to reduce the boilerplate, minimize mistakes, and really spend more time focusing on importance: to build efficient model and extract valuable insights from your data.
Matthew Mayo For,,,,,,,,,, for,, for,,,, for,,,, for,,, for,,,, for,,,, for,,,, for,,, for,,, for,,, for,,, for,,,, for,,, for,,, for,,,, for,,, for,,,, for,,, for,,, for,,,, for,,, for,,, for,,,, for,,, for,,,, for,,, for,,,, for,,, for,,,, for,,, for,,,, for,,,, for,,,, for,,,, for,,,, for,,,, for,,,, for,,, for,,, for,,, for,,, for,,,,, for,,,, for,,,, for,,,, for,, for,.@MattMayo13) Computer science is a graduate diploma in master’s degree and data mining. As the Managing Editor of Kdnuggets & StatologyAnd supporters in the editor Machine specializes in learningMatthew aims to make complex data science concepts accessible. Its professional interests include natural language processing, language models, machine learning algorithms, and the search for emerging AI. He is driven by a mission to democratic knowledge in the data science community. Matthew has been coding since the age of 6.