Prescriptive Analytics and Explainable AI (PGE M1)

<p>High-accuracy estimation of the target outcome (A)</p> Signup and view all the answers

<p>Improving day-to-day operations (C)</p> Signup and view all the answers

<p>Supervised machine learning (A)</p> Signup and view all the answers

<p>The number of chronic conditions (C)</p> Signup and view all the answers

<p>4,406 (C)</p> Signup and view all the answers

<p>Developers of AI building models (C)</p> Signup and view all the answers

<p>Explaining how the AI reflects a faithful representation of real-world scenarios (A)</p> Signup and view all the answers

<p>Technical Complexity (B)</p> Signup and view all the answers

<p>They verify the model’s real-world correspondence and may act as users or managers. (A)</p> Signup and view all the answers

<p>Legal compliance and user impact (A)</p> Signup and view all the answers

<p>External regulators inspecting compliance (C)</p> Signup and view all the answers

<p>Offering practical recommendations based on AI results (A)</p> Signup and view all the answers

<p>Realistic Representation (D)</p> Signup and view all the answers

<p>Predictions are more accurate (B)</p> Signup and view all the answers

<p>To ensure the reliability of the model (D)</p> Signup and view all the answers

<p>The model may become unreliable over time (C)</p> Signup and view all the answers

<p>Meaningful explanations for managerial action (D)</p> Signup and view all the answers

<p>Estimating effects of various actions on target variables (D)</p> Signup and view all the answers

<p>Explanations help to validate accurate predictions. (C)</p> Signup and view all the answers

<p>They guide the analysis to suggest intervention priorities. (B)</p> Signup and view all the answers

<p>Total influence on the values of the concept (C)</p> Signup and view all the answers

<p>Shape the concept to create desired changes (C)</p> Signup and view all the answers

<p>Descriptive analytics (A)</p> Signup and view all the answers

<p>Measure and observe the concept's values (A)</p> Signup and view all the answers

<p>To suggest actions for shaping the future (C)</p> Signup and view all the answers

<p>Providing insights into model biases and fairness (C)</p> Signup and view all the answers

<p>Understanding of factors influencing predictions (A)</p> Signup and view all the answers

<p>It emphasizes analysis results with the greatest impact potential (B)</p> Signup and view all the answers

<p>It explains how AI results can influence human decision-making. (B)</p> Signup and view all the answers

<p>The risks of AI models exposing private information. (D)</p> Signup and view all the answers

<p>When they lack control or have limited control (D)</p> Signup and view all the answers

<p>Illustrating the average predictions and the effects of variables. (D)</p> Signup and view all the answers

<p>Managers focusing on actionable insights for real objectives (C)</p> Signup and view all the answers

<p>More ethical models may sacrifice some level of accuracy. (B)</p> Signup and view all the answers

<p>Interactive explanations for decision support (D)</p> Signup and view all the answers

<p>Understanding how variables influence one another for decision-making. (D)</p> Signup and view all the answers

<p>Every project must have at least one distinct Ultimate concept. (B)</p> Signup and view all the answers

<p>It is studied and confirmed to affect the Ultimate positively or negatively. (D)</p> Signup and view all the answers

<p>They should periodically verify if they still classify as Not Relevant. (D)</p> Signup and view all the answers

<p>Concepts can be designated as high, low, or no control. (B)</p> Signup and view all the answers

<p>To indicate which concepts may affect the Ultimate concept. (C)</p> Signup and view all the answers

<p>Managers should actively shape the concept to their advantage. (D)</p> Signup and view all the answers

<p>It refers to the discernable importance of a concept in affecting outcomes. (B)</p> Signup and view all the answers

<p>Perspective (B)</p> Signup and view all the answers

Analyzing past data to predict future outcomes, assuming the future will resemble the past.

The specific result or event being predicted.

The variables used to predict the target outcome.

A machine learning method primarily used for predictive analytics.

Making highly accurate predictions about a target outcome.

Software that automatically selects and trains machine learning models.

To organize tasks or activities by importance and urgency.

A standard or point of reference for comparison.

Predicting future outcomes and suggesting actions to improve them, focusing on understanding the relationships between input factors and target outcomes.

Making machine learning models understandable by humans, providing insights into how they work and why they produce certain predictions.

Accurate predictions are valuable because they can provide insights into future trends and help make informed decisions.

Without explanations, accurate predictions can be limited in their usefulness because they lack transparency and context.

Meaningful explanations enable managers to understand the relationships between variables and target outcomes, guiding them to make informed decisions.

Simulations allow us to test different scenarios and predict the effects of actions based on accurate predictive models.

The goal of prescriptive analytics is not just to predict the future, but to actively intervene to improve target outcomes.

Prescriptive analytics aims to improve future outcomes by understanding and influencing factors, while predictive analytics focuses solely on predicting future outcomes.

XAI that ensures AI aligns with human values like fairness, privacy, and freedom.

XAI highlighting biases in training data and results to ensure impartial AI outcomes.

XAI addressing the risk of AI models unintentionally revealing private information.

XAI explaining how AI results can guide human decisions to achieve specific goals.

XAI identifying cause-and-effect relationships between variables to understand the reasoning behind AI's recommendations.

XAI providing interactive tools to allow users to simulate and compare different scenarios based on AI's insights.

Visualizations in XAI that demonstrate the impact of each input variable on the AI model's predictions.

Represents how much impact each input variable has on the AI model's predictions, with a stronger effect the further away from the line.

Individuals or groups impacted by AI decisions and involved in its development and regulation.

Explaining how AI arrives at results in a human-understandable way, focusing on the 'why' rather than the technical details.

Ensuring AI's outputs accurately reflect real-world scenarios and can be reliably applied in different contexts.

XAI techniques that make the internal decision-making process of AI models clear and understandable.

XAI techniques that ensure AI models accurately represent the real-world scenarios they are designed to handle.

The degree to which users can rely on AI model outputs to be accurate and reliable.

Experts in specific fields who validate that AI models match real-world scenarios and can be applied effectively.

The ability of different machine learning models to accurately reflect real-world scenarios, with some models being more suited than others.

The most important concept in a project, directly impacting the desired outcome. It's the target or label in supervised learning.

A concept that directly influences the Ultimate Concept and has been confirmed through analysis.

A concept that doesn't influence the Ultimate Concept, confirmed through analysis. It's valuable to identify these to save resources.

A concept that can be manipulated or influenced by managers to impact the Ultimate Concept.

A concept that cannot be directly controlled by managers, but they can still observe and adapt.

Classifications of concepts based on their relevance and controllability, guiding managers on how to best utilize them.

The process of studying concepts to determine their relevance, controllability, and impact on the Ultimate Concept.

The periodic re-evaluation of concepts previously deemed Not Relevant to ensure their continued irrelevance.

The degree to which managers can influence a concept's values. It depends on the relationship between the concept and managerial actions.

Managers have a significant or complete influence on a concept's values. They can directly shape the concept to achieve their desired outcomes.

Managers have some influence on a concept's values, but external factors play a major role. Their actions have a limited impact.

Managers have no influence whatsoever on a concept's values. External, uncontrollable factors determine its course.

The way managers should respond to different levels of control over a concept.

Managers with high control are responsible for shaping the concept to achieve desired outcomes.

Managers with low or no control should focus on measuring and observing the concept, and identifying appropriate responses to its changes.

Managers should proactively identify and implement effective responses to changes in observed values of concepts.