Auto labels · Collapsible communities · Undo/redo · PCA
UPDATE
Start with the Words That Matter.
Fold Complex Groups into a Single Dot.
Even in a keyword network with hundreds of nodes, you can start with the keywords that matter most.
Fold a complex network into groups to see its big structure first.
|
|
|
※ NetMiner 5 is available on Windows and macOS. On Windows, turn on automatic updates in Preferences for a smooth update. (On macOS, we recommend reinstalling separately.)
|
|
|
Automatic node/link label
Auto Hide Small Labels shows or hides labels automatically depending on your zoom level. Link label size to node size (🔗) and the labels of the most important nodes appear first. You can also use Show Top (%) to display only the labels with the highest values.
|
|
|
Treat a community as one group
In a clustered map, collapse a community (group) into a single dot and expand only the ones you want to look at. A collapsed group shows its name and the number of nodes it contains.
|
|
|
Undo and redo
Undo and redo your actions on the canvas, up to 10 steps, with Ctrl+Z and Ctrl+Y.
|
|
|
PCA (Principal Component Analysis)
- Machine Learning > Dimension Reduction > PCA
- Summarize many variables into a few principal components to reduce dimensionality.
|
|
|
Confusion Matrix
- Machine Learning > Evaluation > Confusion Matrix
- Compare a classification model's predicted values with the actual values to see misclassification patterns and performance metrics.
|
|
|
Document attributes preserved in word extraction results
- When you add the Words-Documents Network to the dataset with Add to Dataset, the document attributes of your original data (e.g., date, category) are now saved along with it.
- Menu: Preprocess > Text > Tokenizer, Text > Word Statistics
|
|
|
BERTopic, BERTrend: improved extraction of topic keywords (Topic Info)
-
Fixed an issue where, with Tokenizer checked, 2-grams not in the dictionary were extracted as topic keywords. Even when morphological analysis results were used, 2-grams formed by combining adjacent words (e.g., "analysis analysis") were being generated and extracted as representative words.
-
With this update, the representative words for each topic and the Performance (Topic Coherence) may change. The number of documents per topic and each document's topic number stay the same. (If Target Topic Count is checked, these results may change as well.) See the release notes for details.
|
|
|
We’ve also fixed several known bugs. |
|
|
NEWS
Technical support for NetMiner 4 will end on December 31, 2026.
However, you'll still be able to use and renew NetMiner 4 after that date, and our support team will still be there if you have questions. |
|
|
| Category |
NetMiner 4 |
NetMiner 5 |
| Result interpretation |
None |
AI Assistant
Helps interpret analysis results |
| Word extraction |
Rule/ML-based |
Latest deep learning models |
| Custom dictionary suggestions |
None |
Suggests synonyms, compound nouns, stopwords |
| Topic modeling |
LDA only |
LDA + BERTopic + BERTrend |
| Topic model evaluation |
Topic coherence |
Coherence + perplexity |
| Sentiment analysis |
None |
Deep learning-based positive/negative classification |
| Network formation testing (ERGM) |
Limited MLE |
Standard MCMLE estimation |
| Statistical significance testing |
Permutation-based only |
Permutation + rank-based nonparametric tests |
| Prediction & classification (ML) |
None |
Supported |
| Explainable AI (XAI) |
None |
Feature importance, per-observation contributions (SHAP) |
| Graph-based prediction (GNN) |
None |
Supported |
| Network visualization & charts |
Supported |
Supported |
| Supported OS |
Windows only |
Windows + macOS |
| System requirements & performance |
Requires high RAM |
Optimized RAM usage for large datasets, multiprocessing support |
|
|
-
Existing NetMiner 4 files are fully compatible with NetMiner 5, with the exception of process logs.
-
Even when a menu name stays the same, the underlying method behind it may have improved (for example, Text Analysis > Word Extraction or LDA) — in these cases, results can differ from what you got in NetMiner 4.
-
To check whether a NetMiner 4 feature you rely on is available in NetMiner 5, please see the online user manual. If you have questions about a specific menu or feature, feel free to reach out to us.
|
|
|
NetMiner is a social network analysis / graph analytics platform
and provides all the features to support your data analysis workflow
|
|
|
ⓒ CYRAM Inc. | 253, Pangyo-ro, Bundang-gu, Seongnam-si, Gyeonggi-do, Republic of Korea
|
|
|
|
|
