Programming · PRC-34

Machine Learning and Deep Learning with KNIME

Machine Learning and Deep Learning with KNIME is a course in building analytics as visual workflows in KNIME Analytics Platform 5, from data preparation and evaluating classification and clustering models to neural networks, CNNs and LSTMs. It suits data analysts, data scientists and teachers who want to work with machine learning without writing much code.

Updated 8 Oct 2026

From 12,150 THB / person 13,500 −10% excl. VAT 7% · group rates available

Corporate training quote See open rounds

PDFDownload the course outline

Duration18 hours · 3 days
FormatOnsite / live online
Next round28–30 Dec 26
CertificateIncluded

Course overview

Many organisations already hold plenty of data but cannot turn it into forecasting models or customer segments because their teams are not comfortable writing code. KNIME Analytics Platform is an open source tool that builds analytics by connecting nodes into a workflow, so every step from reading the data to the final result stays visible. That makes it a good fit for beginners, teachers and data teams who need work that can be checked later and handed over.

This course takes learners through the current KNIME Analytics Platform 5, from preparing data with nodes end to end to building classification, regression, association rule and clustering models and evaluating and tuning them properly. It continues with deep learning, covering neural networks, CNNs and LSTMs through the Keras nodes, and shows how to use K-AI, the Python Script node and LLMs inside a workflow. It closes with a project that separates the training workflow from the scoring workflow, as is done in real work. (3 days, 6 hours per day, 18 hours in total, Beginner level.)

What you’ll gain

Install KNIME Analytics Platform 5 and build analytics workflows on your own
Import, explore and prepare data with nodes, including converting, filtering, joining, aggregating and normalising
Build visualisations, components and interactive data apps to communicate results
Build classification and regression models and evaluate them with cross-validation and ROC
Tune hyperparameters and compare several algorithms systematically
Apply association rules and clustering to find products bought together and segment customers
Build neural network, CNN and LSTM deep learning models with the Keras nodes
Use K-AI, the Python Script node and LLMs in workflows, and separate a scoring workflow for production use

Who this course is for

Data analysts and data scientists who want a visual workflow tool whose work is easy to check and hand over
Data engineers and developers who need to prepare data and build models without coding every step
Business analysts and staff who want to get started with machine learning and deep learning
Lecturers, trainers and teachers who want to demonstrate the analytics process clearly
Teams that want a shared standard for working with data

Prerequisites

Comfortable with computers and Excel, and familiar with tables of rows and columns
No programming experience needed; basic Python helps you go further
Basic statistics such as averages and spread will help you learn faster
A laptop with 16 GB of RAM or more and permission to install software

Curriculum

Course Details

This course is the 2026 edition of Machine Learning and Deep Learning with KNIME, updated for KNIME Analytics Platform 5. It runs for 3 days, 6 hours per day (18 hours in total, 09:00-16:00), as lectures with a lab in every topic. Beginner level, with no programming experience required. Day one focuses on data preparation and visualisation. Day two covers machine learning, including classification, regression, association rules and clustering, together with model evaluation and tuning. Day three covers deep learning with the Keras nodes, using LLMs in workflows and a capstone that puts a model into use. All tools are open source or free, and the deep learning labs run on a CPU. Learners take home a lab guide, KNIME workflows and sample datasets for every lab.

Day 1 KNIME 5 and Data Preparation

Section 1: Lab: Getting Started with KNIME Analytics Platform 5

What KNIME is: node-based visual workflows and low-code data science
Install KNIME 5 and choose between the LTS and the standard release
The new interface: Node Repository, Description, Node Monitor and Space Explorer
Configure and execute nodes, read node status, and save and export workflows
Install extensions and find example workflows on the KNIME Community Hub

Section 2: Lab: Importing and Exploring Data

Read data with CSV Reader, Excel Reader and the DB nodes for databases
Check the schema, data types and basic statistics of each column
Find missing values, duplicates, malformed values and outliers
Lab: explore a customer dataset and summarise its data quality issues

Section 3: End-to-End Data Preparation with Nodes

Convert types, use String Manipulation and create new columns with the Expression node
Missing Value, Rule Engine, Row Filter and Column Filter
Joiner, Concatenate, GroupBy and Pivot to combine and summarise data
Binning, Normalizer and One to Many to prepare features
Lab: build a workflow that takes a raw file all the way to model-ready data

Section 4: Visualisation, Components and Data Apps

Bar charts, scatter plots, box plots and heatmaps to see patterns in data
Group nodes into metanodes and components to keep workflows readable and reusable
Build interactive views with component views and open them as a data app
Lab: a customer exploration data app with selectable filters

Section 5: Lab: K-AI and Python in KNIME

K-AI, the AI assistant, in both Q&A mode and build mode that adds nodes for you
What data leaves your machine when you use K-AI, and why it matters
The Python Script node with the knime.scripting.io module and the bundled environment
Lab: use pandas to extend a workflow where built-in nodes fall short

Day 2 Machine Learning with KNIME

Section 6: Machine Learning Concepts and Splitting Data

Supervised versus unsupervised learning and the steps of a machine learning project
Stratified splits with Partitioning and cross-validation with X-Partitioner
Data leakage and keeping data preparation separate from the test set
Class imbalance and handling it with SMOTE or class weights

Section 7: Lab: Building Classification Models

Decision Tree, Naive Bayes and k-Nearest Neighbor
Logistic Regression and Support Vector Machine
Ensemble models: Random Forest, Gradient Boosted Trees and XGBoost
Lab: predict customers who are likely to churn

Section 8: Evaluating and Tuning Models

Confusion matrix, accuracy, precision, recall and F1-score with the Scorer
ROC curves, AUC and choosing a threshold that matches business costs
Find the best hyperparameters with the Parameter Optimization Loop
Feature selection and explaining models with feature importance
Lab: compare several algorithms in a single table

Section 9: Regression and Association Rules

Linear Regression and Regression Tree, measured with MAE, RMSE and R²
Market basket analysis: support, confidence and lift
Build rules with Association Rule Learner and turn them into actions
Lab: forecast sales and find products that customers often buy together

Section 10: Workshop: Clustering and Customer Segmentation

k-Means and choosing the number of clusters with the elbow and silhouette methods
Hierarchical Clustering and DBSCAN for data with complex shapes
Compare the methods and describe the profile of each cluster
Workshop: segment customers and summarise marketing recommendations

Day 3 Deep Learning, LLMs and Putting Models to Work

Section 11: Lab: Deep Learning Basics with the Keras Nodes

Neural networks, layers, activation functions, loss and optimisers
Overfitting, dropout and stopping training when results stop improving (early stopping)
Set up the Conda environment for the KNIME Deep Learning Integration
Keras layer nodes, Keras Network Learner and Keras Network Executor
Lab: a neural network for classifying tabular data

Section 12: Lab: CNNs for Image Classification

How convolution, pooling and feature maps work
Read and prepare images with the KNIME Image Processing extension
Transfer learning from a pre-trained model to cut training time and data needs
Lab: classify sample product images and measure results with a confusion matrix

Section 13: LSTMs and Time Series Data

The nature of time series data and creating lag features with Lag Column
Split data by time rather than at random so future data does not leak into the model
LSTM layers for predicting the next value in a sequence
Lab: forecast daily sales with an LSTM and compare it with a baseline model

Section 14: Python Deep Learning and LLMs in Workflows

Train models with Keras 3 or PyTorch in the Python Script node when built-in nodes are not enough
KNIME AI Extension: connect to a local LLM through Ollama and use the LLM Prompter
Classify and summarise customer feedback text with an LLM inside a workflow
Vector stores and basic RAG, with care for personal data
Lab: analyse customer reviews with an LLM alongside an existing model

Section 15: Workshop: Putting Models to Work and Capstone

Separate the training workflow from the scoring workflow
Save models, run workflows in batch and share them through KNIME Hub
Capstone: choose a classification, clustering or deep learning problem from the sample data
Present the results, review them together and wrap up with a checklist for using models in real work

Schedule & training options

For individuals — public rounds

28–30December

Classroom28–30 Dec 26

20 seats left

23–25February

Classroom23–25 Feb 27

20 seats left

20–22April

Classroom20–22 Apr 27

20 seats left

8–10June

Classroom8–10 Jun 27

20 seats left

24–26August

Classroom24–26 Aug 27

20 seats left

26–28October

Classroom26–28 Oct 27

20 seats left

28–30December

Classroom28–30 Dec 27

20 seats left

For organisations — in-house / private

Tailor the content to your team’s tools and projects
Your dates, at your office or live online
Quotation with tax ID for procurement

Corporate training quote

Instructors

Sanitwong KamonpakornInstructor

Frequently asked questions

What is KNIME, and how do visual workflows work?

KNIME Analytics Platform is open source software for analysing data and building models. Each step, such as reading a file, filtering data or training a model, is a node, and nodes are connected into a workflow. You configure each node in a dialog and can inspect the output at every step, so the work is easier to check, change and hand over than a long script. You can still add a Python Script node when you need more flexibility.

How do precision, recall and F1-score differ?

Precision tells you how many of the cases the model predicts as positive are actually positive. Recall tells you how many of all the real positive cases the model catches. F1-score is the harmonic mean of the two and is useful on imbalanced data. Which one matters depends on the cost of each mistake; churn screening, for example, usually favours recall. In the course you calculate these with the Scorer node and use ROC curves to choose a threshold.

What kinds of tasks are CNNs and LSTMs suited to?

CNNs (convolutional neural networks) suit image data because they learn textures, edges and shapes from neighbouring pixels, so they are used for image classification and object detection. LSTMs are networks that remember context across a sequence, which suits data ordered in time, such as daily sales or sensor readings. In the course you build both with the Keras nodes in KNIME and always compare them with a simple baseline model.

Who is Machine Learning and Deep Learning with KNIME for, and what background do I need?

It suits data analysts, data scientists, data engineers, business analysts and teachers who want to build models with a visual workflow tool. No programming experience is needed; you only need to be comfortable with computers, Excel and tables of data. Basic statistics or some Python will help you go further. Please bring a laptop with 16 GB of RAM or more on which you can install software.

What can I do after completing Machine Learning and Deep Learning with KNIME?

You will be able to build complete data preparation workflows, build and compare classification, regression and clustering models, evaluate them with cross-validation and ROC, and tune hyperparameters. You can start deep learning work with neural networks, CNNs and LSTMs, use K-AI, Python Script and LLMs in your workflows, and set up a separate workflow that scores new data. You also take home workflows for every lab.

Can Machine Learning and Deep Learning with KNIME be run in-house for our team?

Yes. IT Genius Institute runs in-house training for organisations. Lab examples and problems can be adapted to your team's data and work, such as sales forecasting, customer segmentation or text analysis, and the emphasis can lean towards machine learning or deep learning as needed. Contact the institute to request a quotation.

Articles are published in Thai.