ME&A_​2025&26_​group_​AP_​KNIME (1)

These nodes convert categorical variables into numeric format by replacing text values with numerical codes.

This node converts values stored as text (string) into numeric formats (integer/double), enabling their use in calculations and data analysis.

Identifies extreme values using the IQR method and caps them at acceptable limits to reduce their impact without removing observations.

Missing values were handled to ensure data completeness and avoid errors during analysis.

Numeric Outliers: Treated to reduce their impact on clustering and predictive models.

Missing Values: Replaced to ensure a complete dataset.

Income → Median: Used because it is more robust to outliers.

Other Numeric Variables → Mean: Used to preserve the overall distribution.

String Variables → Most Frequent Value: Used to maintain consistency in categorical data.

This node applies logical rules to transform and recode variables, preparing the data for analysis.

This node selects the variables relevant for the analysis by removing unnecessary columns.

This node scales all variables to a common range, enabling fair comparisons during clustering.

Table manipulator: This node reorganizes the table structure by renaming, reordering, or modifying columns as required for the analysis.

RowID: This node creates or modifies row identifiers, ensuring that each observation has a unique ID.

Statistics: This node generates descriptive statistics to summarize the main characteristics of the data.

Histograms: This node visualizes the distribution of variables, helping to identify patterns, skewness, and potential outliers.

K-Means: This node groups customers into clusters based on similarities in their characteristics.

Silhouette Coefficient: This node evaluates the quality of the clustering by measuring how well each observation fits within its assigned cluster.

GroupBy: This node calculates summary statistics for each cluster, enabling cluster profiling and comparison.

Color Manager: This node assigns colours to clusters to improve visual interpretation.

Scatter Plot: This node visualizes the clusters, making it easier to identify patterns and separation between groups.

Desnormalizer: This node restores the variables to their original scale, making the results easier to interpret.

Table Partitioner: Splits the dataset into training and testing sets for model validation.

Decision Tree Learner: Builds a decision tree model using the training data.

Decision Tree Predictor: Applies the trained model to generate predictions.

Scorer: Evaluates the predictive performance of the model using classification metrics.

ROC Curve: Assesses the model's ability to distinguish between classes.

Lift Chart: Measures the improvement of the model over random selection.

Linear Correlation: Identifies relationships between numerical variables.

Correlation Filter: Removes highly correlated variables to reduce redundancy and improve model performance.

RProp MLP Learner: Builds a Multi-Layer Perceptron (Neural Network) model using the training data.

MultiLayerPerceptron Predictor: Applies the trained neural network to generate predictions on new data.

Excel Writer: Exports the results and model outputs to an Excel file for further analysis and reporting.

Linear Correlation: Identifies relationships between numerical variables.

Heatmap: Visualizes correlations between variables through a colour-coded matrix.

Statistics: Provides descriptive statistics to summarize the dataset.

Histogram: Displays the distribution of a variable and helps identify patterns and outliers.

Scatter Plot: Visualizes the relationship between two variables and helps detect trends or clusters.

Bar Chart: Displays the frequency or distribution of categories, making it easier to compare values across groups.