Exercices - Intelligence artificielle et données
20 documents à télécharger gratuitement
Exercices de intelligence artificielle et données partagés par des étudiants et des enseignants. Thèmes couverts : machine learning, apprentissage automatique, big data, data science, fouille de données.
The document provides solutions to a series of exercises focused on data mining concepts such as centroid calculation, Euclidean distance, and inertia determination relative to point clouds. Methodologies such as Min-Max normalization, Z-score normalization, and decimal scaling are detailed, with numerical examples provided. The core mathematical computations and formulas used in data processing are demonstrated, emphasizing practical applications in data normalization and cluster analysis. The findings include specific worked-out values and examples indicative of their real-world usage in...
This document details the procedure for creating an Extract, Transform, Load (ETL) solution using SQL Server Integration Services (SSIS). It includes instructions for restoring backup files, utilizing SQL Server Management Studio, and configuring a local environment. The methodology emphasizes the preparation of databases and execution of sequential tasks across multiple exercises. Findings suggest a step-by-step approach for successful ETL implementation within the (local) SQL Server environment.
The document provides a thorough exploration of probabilistic classification methods, including the majority rule, maximum likelihood rule, and Bayesian rule. It explains key concepts such as how probabilities are defined within datasets and highlights the differences between these methods in terms of accuracy and applicability. The Bayes rule is specifically derived and compared to other methods, presenting its scenarios where it simplifies to maximum likelihood. Finally, error computations for each classification method are detailed to demonstrate their effectiveness.
The document focuses on probabilistic approaches in classification, emphasizing the use of Bayesian decision rules. It evaluates weather predictions using barometric data, exploring how prior and conditional probabilities combine to predict outcomes. Additionally, it addresses the classification of a population based on logical attributes, examining three methods: majority rule, maximum likelihood, and Bayesian classification. The study highlights the strengths and weaknesses of each method, employing mathematical and probabilistic techniques to calculate prediction errors and assess classi...
This document examines probabilistic approaches in the context of machine learning, focusing on Bayesian decision-making. It introduces a weather prediction problem based on barometer readings using supervised learning, where Bayes' theorem is applied to derive conditional probabilities. The methodology includes calculation of posterior probabilities to classify observations and assess the reliability of predictions. The exercise concludes with a demonstration of the Bayesian classification process, reaching a prediction reliability of 67% when forecasting rain based on barometer observations.
This document focuses on perceptron algorithms for learning tasks. Exercise 1 addresses the classification of linearly separable data in a 2D space using a decision boundary defined by y=x/2, and provides a sample dataset for step-by-step algorithm training. Exercise 2 involves designing a perceptron to determine the parity of digits displayed by a seven-segment numeral, enriching practical understanding of binary classification. Initial weights are defined at zero, and examples alternate between positive and negative classes during training. The visual traces of the algorithm illustrate th...
This document introduces and applies correspondence analysis (CA) to multidimensional contingency tables. The methodology involves partitioning data into classes, visualizing their positions on principal component graphs, and exploring hierarchical relationships through ultrametric distances. Findings include detailed correspondence patterns and profiles derived from respondent data, along with specific metrics such as eigenvalues and chi-squared statistics, which provide insights into variable relationships.
This document covers concepts related to multidimensional analysis, including class partitioning and ultrametric distance evaluation. It presents an exercise involving brand name selection based on customer preferences using multidimensional data analysis methods. The results are summarized in tables and graphical representations, highlighting frequency distributions.
This document details a set of exercises aimed at using Apache Spark's RDD functionalities for processing large datasets. The tasks include loading datasets, filtering data based on specific criteria, working with partitions, performing transformations using map and reduceByKey, counting word occurrences in multiple files, and combining results through joins. The methodology emphasizes practical implementation of Spark's RDD operations for handling logs, purchase data, and textual analysis while utilizing caching for optimization.
This document outlines an exam for the Master Professional in Business Intelligence at the Université Virtuelle de Tunis. It consists of exercises focusing on data modeling and integration for a chain of stores specializing in newspapers and magazines. The exam assesses the ability to create star and snowflake schemas, along with SQL query formulation and analysis of sales data.
This document presents exercises for a Master's level course on data warehousing. It includes tasks on designing data marts to analyze sales data in various contexts, such as disposable dish manufacturing and fast-food restaurants. The exercises also involve creating models in star and snowflake schemas.
This document explores the application of fuzzy logic in two main scenarios: quality control in manufacturing and temperature regulation of buildings. The first exercise involves constructing a fuzzy logic-based decision-making system for smartphone quality based on weight and length, utilizing trapezoidal and triangular membership functions. The second exercise designs a temperature regulation system for buildings using fuzzy controllers, with inputs from internal and external temperature sensors, and varying power outputs determined by predefined linguistic terms and membership functions....
Ce document contient des exercices sur l'inférence floue dans le cadre d'une matière d'intelligence artificielle. L'exercice porte sur l'application de la méthode Max-Min et l'évaluation des températures à l'aide de règles d'inférences. Il inclut également des représentations graphiques pour illustrer les fonctions d'appartenance.
This document outlines coursework focused on fuzzy logic, specifically applied to quality control in manufacturing and temperature regulation in buildings. It includes exercises that require the use of fuzzy sets and decision-making processes. Students are tasked with graphical representation of fuzzy subsets and formalization of membership functions.
This document focuses on the perceptron algorithm's application to a linearly separable dataset and a parity problem. In the first exercise, a dataset with binary targets defined by a linear boundary (y = x/2) is used for training, starting with weights initialized to zero. Examples are alternately presented until all are correctly classified. The second exercise extends the perceptron to classify parity using a seven-segment display of digits. Trace outputs highlight algorithmic iterations and updates during learning.
This document discusses the implementation and evaluation of linear regression models using Python and the Scikit-learn library. The methodology covers key processes such as minimizing the residual sum of squares (RSS) to optimize the model's efficiency, calculating critical metrics including variance, covariance, and parameters α and β, and exploring methods to avoid overfitting. Findings also include the use of R-squared to validate model predictions, yielding a determination coefficient of 0.66.
This document introduces linear regression as a fundamental machine learning model, applied to predict the relationship between pizza size and its price using Python libraries such as numpy and matplotlib for data generation and visualization. Scikit-learn's LinearRegression model is employed for both training and prediction, demonstrating a linear relationship validated through plotted data and model equations. The methodology highlights step-by-step processes of creating data arrays, visualizing trends with graphs, and applying fit-predict functions for deriving the regression equation.
The document outlines an estimation process for the current year's consumption of an unspecified article based on the previous year’s monthly data and related percentage adjustments (Rn values). The data provided includes explicit monthly consumption figures and corresponding variation percentages. The methodology involves predicting future values by applying these percentages to the prior year's consumption data, with a specific rounding requirement for the final predicted values. This serves as a mathematical and statistical exercise in forecasting based on historical data trends.
The document contains exercises focusing on decision-making models in various scenarios, including production decisions in response to milk shortages, weather forecasting for carrying an umbrella, and decision-making at a tennis tournament. Each exercise requires calculations of expected utility and comparisons of different decision models.
The document consists of exercises that explore decision-making models using graphical representations. It includes scenarios where individuals must make strategic decisions based on uncertain outcomes, using concepts from probability and utility theory. The exercises guide the reader through creating decision trees and Bayesian networks for optimization.