Time Series Forecasting with Python

تنبّأ بما سيأتي باستخدام Python. على مدى 3 أيام تستكشف السلاسل الزمنية بـ pandas وNumPy

3 أياممتوسط6 وحدات

نظرة عامة

تنبّأ بما سيأتي باستخدام Python. على مدى 3 أيام تستكشف السلاسل الزمنية بـ pandas وNumPy، وتفكّك الخطأ والاتجاه والموسمية بـ statsmodels، وتختبر الاستقرارية، وتبني نماذج ARIMA وSARIMA وSARIMAX، وتدرّب الشبكات العصبية المتكرّرة في TensorFlow، وتنمذج الموسمية بسرعة باستخدام Prophet، وتتحقّق من التنبؤات ببيانات الاختبار.

المنهاج

المحتوى التقني أدناه معروضٌ باللغة الإنجليزية كما ورد في منهاج الدورة، حفاظاً على دقّة المصطلحات.

6 وحدات

Introduction to time series forecasting and predictive analytics
Exploring time series data with pandas and NumPy
Importing and preparing datasets (import pandas as pd, import numpy as np)
Visualizing trends, seasonal patterns, and moving averages
Overview of forecasting involves and common real-world use cases

Using statsmodels for statistical time series analysis
Decomposing time series with error-trend-seasonality models
Identifying stationary vs. non-stationary datasets
Detecting lag structures with ACF and PACF plots
Understanding model uses for autoregressive and MA processes

Forecasting with ARIMA and SARIMA models
Developing SARIMAX models to include exogenous variables
Working with multiple variables and multiple models
Evaluating model performance with RMSE and forecast accuracy
Forecasting time series data and making data-driven decisions

Introduction to deep learning for forecasting
Building and training Recurrent Neural Networks using TensorFlow
Using neural networks for model training and prediction
Tuning hyperparameters and tracking model results
Predicting the future based on historical data

Overview of the Prophet library and its applications
Modeling trends and seasonality with minimal tuning
Building scalable forecasting models using Prophet
Comparing Prophet to classical methods and machine learning models
Forecast future values based on real datasets

Applying random forest and other machine learning models to time series
Performing exploratory data analysis before model fitting
Using test data to validate predictions
Making accurate forecasts with real-world datasets like stock price trends
Analyzing results from statistical models, predictive tools, and neural architectures
Using linear combination techniques in advanced models

لمن هذه الدورة

محلّلو البيانات ومستخدمو Python الذين يتنبّؤون بالطلب أو المبيعات أو حركة الزوّار أو غيرها من القيم عبر الزمن.