Identification of Potentially Risky Publications

Identification of Potentially Risky Publications

Risky Mentions

In the Semantrum system, the system tone of voice toward the monitored object is automatically identified. To make the analysis of negative tone of voice more detailed, we have added additional categories for publications with negative sentiment.

The system now automatically identifies the following categories of risky mentions:
  1. Financial Operations: These include (but are not limited to) financial crimes, financial schemes, fraudulent financial activities, tax fraud, deceptive investment projects, financial manipulations and market manipulations, bank fraud, excessive use of credit, and debt traps.
  2. Corruption Risks: Corruption, bribery, abuse of office for personal gain.
  3. Antimonopoly legislation: Violations in the field of antitrust laws and unfair competition.
  4. Ecology: Environmental crimes.
  5. Customs Legislation: Smuggling, counterfeit goods, drugs, weapons, human trafficking.
  6. Cyber Threats, Personal Data.
  7. Labor Legislation: Violations in labor relations and employee rights.
  8. Intellectual Property.
  9. Ethical Norms: Manipulation of assessments, judgments by officials or company representatives, violation of business principles, standards, and integrity rules.
  10. Sanctions, National Security: Anti-state activities, violations of sanction restrictions, inclusion in sanction lists, collaboration with sanctioned companies/countries/persons, links to extremist or terrorist groups.
  11. Offenses and Accidents.

Risk categories are determined in the system using AI, specifically through Machine Learning technology, trained on a large amount of manually labeled data by media analysts. These categories are only defined for publications in Ukrainian and English.


Please note:
Publications flagged as risky do not indicate or prove any offenses or risks; they simply draw the user's attention to publications selected by the algorithm using Machine Learning for specific categories. The tone of voice and riskiness of publications are determined separately by ML algorithms. Publications of any tone of voice can be classified as risky.

How to Set Up Risk Detection for Publications

To configure the functionality of identifying risky publications, please contact your support manager or write to us in the system chat.

Displaying Risk Categories in the Publication Feed

If the risk identification feature is enabled in your topics, you will be able to see the identified risk categories on the publication card in the news feed. Hover over the risk icon, and you will see the specific risk category or categories detected in the news.




Filtering by Risky Publications

On the filter panel, you can filter publications by identified risk categories. You can choose one category, publications belonging to several categories simultaneously, publications belonging to at least one of the selected categories, or exclude one or several categories. To manage category selection, use the menu next to the category name. Learn more about the filter management.



Alerts for Risky Publications

To stay informed about important publications that may pose certain risks to your company's or brand's reputation, you can set up alerts for specific risk categories. Select the topics in which risk identification is configured, choose the maximum period for which you have accumulated risk data to cover all available risk categories in the topic, and create a saved query. Learn more about saved queries.



Then, open the notification settings, select the necessary saved query, and configure convenient alert options—email or Telegram, in real-time or on a schedule, with attachments or without. Newsletters (allerts) in Semantrum: overview of functions. How to set up newsletters by saved queries.



Done! You won’t miss important information.

Enjoy using the Semantrum system!


                Learn more about saved queries
                How to set up newsletters by saved queries
                Tone of voice determination

    • Related Articles

    • Tone of voice determination

      What is tone of voice (emotion)? Tone of voice is the emotional connotation of a message, usually containing the author's subjective evaluation of a certain object. It is important to realize that this is not a state of fact, pleasant or not, but the ...

    • How to change the tone of voice of publications

      If you need to change the publications' tone of voice, you can use the Mass Tone Change and Story Tone Change tools. Please note: You can change the tone of a maximum of 3,000 posts at a time. If you select more publications, the system will process ...

    • Automated visual analytics (Charts mode)

      What is "Charts" mode The Charts mode provides access to a quick visual analysis of mentions according to selected filters. The main charts for quantity statistics of mentions are presented here. Several tabs with thematically grouped graphs are ...

    • About Semantrum system: basic features

      Semantrum is an online system for reputation monitoring and analysis. The system searches for your and your competitors' brands, companies, organizations, people and other keywords in a database of data collected from web media, social media, review ...

    • What is a dashboard and what are its features?

      What is a dashboard? A dashboard is a tool for visualizing monitoring results, a set of interactive graphs and indicators customized for one or more topics. Using dashboard makes data analysis much easier, because you get summarized data of your ...