Icon

04_​Anomaly_​Detection_​Solution

Anomaly Detection (Exercise) - Solution

This workflow instructs you to detect anomalies in contracts via the IQR, z-score, and isolation forest techniques. Furthermore, shows you how to customize a dashboard to improve its appearance.

Loading the Contracts 10 contracts as pdf filestransformed into Documentobjects Extract the following information from eachcontract- Date: the date, on which the contract was signed - Contract ID: Each document contains a 7-charecter ID (e.g. C000058)- Payment Amount: Each document contains apayment amount Customize a dashboard Detect and visualize outliers based on the IQR Detect outliers based on the z-score Session 4Exercise - Detect outliers and customize a dashboardSummary:In this exercise, we will use different techniques to detect and visualize outliers andapply best practices for dashboarding. Instructions:1) Execute the preprepared workflow to access some contracts and extract theirdates, IDs, and payment amounts2) First, use the following nodes to detect outliers in the payment amount based onthe IQR:-Numeric Outliers with k=1.5-Box Plot3) Second, complete the following steps to detect outliers in the same columnbased on the z-score:-Normalize the data with the Normalizer node-Use the Math Formula and Rule Engine nodes to flag payments with normalizedabsolute values greater than 3-Denormalize the data with the Denormalizer node4) Finally, polish a dashboard visualizing outliers as follows:- Execute the H2O Isolation Forest and Visualize Outliers components. The formerperforms the outlier detection using the isolation forest algorithm, and the lattervisualizes the results.-Change the layout of the composite view for a more optimal dashboarddistribution. There is no one right solution but keep in that the most importantinformation is positioned in the top left corner and in the middle.-Increase the size of the title using the html formatting in the Text Output Widgetnode-Change the font of category labels yes/no in the bar chart with the CSS Editornode. You can use the following code:text.knime-tick-label { font-size: 14px; font-weight: bold;} Finding outliersExtracting Contract ID, Date and Payment amounts from the texts using RegExz-score|z| > thr?|z| sign(z)sign(z)20 levels100 treesGetting the text from PDFs Numeric Outliers Extract Date, ContractID and Payment Amount Normalizer Rule Engine Math Formula Denormalizer Math Formula Math Formula Visualize Outliers H2O IsolationForest PDF Parser Box Plot Loading the Contracts 10 contracts as pdf filestransformed into Documentobjects Extract the following information from eachcontract- Date: the date, on which the contract was signed - Contract ID: Each document contains a 7-charecter ID (e.g. C000058)- Payment Amount: Each document contains apayment amount Customize a dashboard Detect and visualize outliers based on the IQR Detect outliers based on the z-score Session 4Exercise - Detect outliers and customize a dashboardSummary:In this exercise, we will use different techniques to detect and visualize outliers andapply best practices for dashboarding. Instructions:1) Execute the preprepared workflow to access some contracts and extract theirdates, IDs, and payment amounts2) First, use the following nodes to detect outliers in the payment amount based onthe IQR:-Numeric Outliers with k=1.5-Box Plot3) Second, complete the following steps to detect outliers in the same columnbased on the z-score:-Normalize the data with the Normalizer node-Use the Math Formula and Rule Engine nodes to flag payments with normalizedabsolute values greater than 3-Denormalize the data with the Denormalizer node4) Finally, polish a dashboard visualizing outliers as follows:- Execute the H2O Isolation Forest and Visualize Outliers components. The formerperforms the outlier detection using the isolation forest algorithm, and the lattervisualizes the results.-Change the layout of the composite view for a more optimal dashboarddistribution. There is no one right solution but keep in that the most importantinformation is positioned in the top left corner and in the middle.-Increase the size of the title using the html formatting in the Text Output Widgetnode-Change the font of category labels yes/no in the bar chart with the CSS Editornode. You can use the following code:text.knime-tick-label { font-size: 14px; font-weight: bold;} Finding outliersExtracting Contract ID, Date and Payment amounts from the texts using RegExz-score|z| > thr?|z| sign(z)sign(z)20 levels100 treesGetting the text from PDFsNumeric Outliers Extract Date, ContractID and Payment Amount Normalizer Rule Engine Math Formula Denormalizer Math Formula Math Formula Visualize Outliers H2O IsolationForest PDF Parser Box Plot

Nodes

Extensions

Links

Download

To use this workflow in KNIME, download it from the below URL and open it in KNIME:

Download Workflow
Created by: dietzc
Created at: 2015-12-08
On NodePit since: 2023-04-06
Last update: 2024-04-26
Created with KNIME version: v4.7.2
Tags: fraud detectionanomaly detectiontext processingpdf documentspdfoutlier detectioncontractseducation

Deploy, schedule, execute, and monitor your KNIME workflows locally, in the cloud or on-premises – with our brand new NodePit Runner.

Try NodePit Runner!