Logistic regression

Polynomial regression

Ridge regression

Linear regression

Linear regression uses a straight line; polynomial regression uses a polynomial equation.

Linear regression uses a quadratic equation; polynomial regression uses a straight line.

Linear regression fits multiple curves; polynomial regression fits a single line.

Linear regression is used for classification; polynomial regression is used for regression.

The training dataset is used solely for testing the model's accuracy.

The training dataset is essential for fitting the regression model and learning the relationship between variables.

The training dataset provides the final predictions for new data points.

The training dataset is irrelevant to the regression model's performance.

Feature layer, decision layer, output layer, gradients.

Input layer, output layer, training data, learning rate.

Input layer, hidden layers, output layer, weights, activation functions.

Input nodes, output nodes, bias terms, loss functions.

Neural networks can be used for predicting structural behavior, optimizing designs, analyzing sensor data, automating construction, and improving project management.

Neural networks can replace traditional materials in building projects.

Neural networks are primarily used for financial forecasting in construction.

Neural networks can enhance aesthetic design choices in architecture.

Overfitting occurs when a neural network generalizes well to new data but fails on training data.

Overfitting is when a neural network performs well on training data but poorly on test data due to excessive learning of noise.

Overfitting is when a model learns too little from the training data and performs poorly overall.

Overfitting happens when a neural network is too simple to capture the underlying patterns in the data.

Activation functions are only used for output layers in networks.

Activation functions are optional and do not affect network performance.

Activation functions are crucial for enabling neural networks to learn complex patterns and introduce non-linearity.

Activation functions simplify the learning process by removing complexity.

Time series models, causal models, qualitative models

Statistical tests, sampling methods, data mining

Predictive models, regression models, descriptive models

Exploratory models, inferential models, experimental models

Time series analysis relies solely on expert opinions for predictions.

Time series analysis focuses on qualitative data rather than quantitative trends.

Time series analysis uses random sampling to create forecasts.

Time series analysis contributes to forecasting by identifying patterns and trends in historical data to predict future values.

To create visualizations of data trends.

To summarize historical data in a report.

To analyze and predict future points in a time series.

To determine the mean of a dataset.

Qualitative forecasting is subjective and opinion-based, while quantitative forecasting is objective and data-driven.

Qualitative forecasting is based on numerical data, while quantitative forecasting is based on personal insights.

Qualitative forecasting relies on statistical models, while quantitative forecasting uses expert opinions.

Qualitative forecasting is focused on historical trends, while quantitative forecasting emphasizes future predictions.

Monte Carlo simulations are used to evaluate material strengths.

Monte Carlo simulations are used to design building layouts.

Monte Carlo simulations are used to assess risks and uncertainties in civil engineering projects.

Monte Carlo simulations are used to calculate construction costs.

Simulation techniques reduce project planning efficiency through excessive data analysis.

Simulation techniques improve project planning by enabling scenario modeling, risk assessment, and resource optimization.

Simulation techniques limit project planning by focusing solely on historical data.

Simulation techniques hinder project planning by complicating decision-making.

Sensitivity analysis is primarily focused on data visualization techniques.

Sensitivity analysis is important for understanding the impact of input changes on model outputs, aiding in risk assessment and decision-making.

Sensitivity analysis helps in determining the exact values of model parameters.

Sensitivity analysis is used to eliminate unnecessary variables in a model.

Stochastic models provide fixed outcomes, whereas deterministic models vary with each run.

Deterministic models include randomness, while stochastic models are entirely consistent.

Deterministic models yield consistent results, while stochastic models account for variability and randomness.

Deterministic models are always random, while stochastic models are predictable.

MATLAB requires external software for any visualization tasks.

MATLAB can visualize simulation results using plotting functions, 3D visualizations, and Simulink.

MATLAB is limited to 2D graphs for all visualizations.

MATLAB can only display text outputs from simulations.

Cross-validation is a method to increase model complexity and reduce accuracy.

Cross-validation is used to train models on the same data repeatedly for better results.

Cross-validation is important in model evaluation because it helps to prevent overfitting, provides a more accurate estimate of model performance, and ensures that the model generalizes well to unseen data.

Cross-validation is a technique that only applies to linear regression models.

Coefficients represent the total value of the independent variable.

Coefficients indicate the correlation between two independent variables.

Coefficients indicate the expected change in the dependent variable for a one-unit change in the independent variable.

Coefficients show the average of all dependent variable values.

Advantages of using MATLAB for predictive modeling include user-friendly interface, built-in functions, extensive visualization tools, support for various algorithms, and efficient handling of large datasets.

Inadequate data handling capabilities

Complex coding requirements

Limited algorithm support