The In-Depth View of the Similarity Analysis provides detailed insights into Original Language Similarity and Translated Language Similarity, individually, at both the document and sentence levels.

1. Original Language Similarity

Original language similarity refers to the degree of similarity detected between the document and other sources in its original language. Users can access insights at the document level to see all instances of similarity throughout the entire document or at the sentence level for a specific sentence.

1.1 Document Level Insights (Original Language Similarity)

At the document level, Inspera Originality provides detailed information for all Original Language Similarity findings throughout the entire document.

The Original Language Similarity percentage represents the overall similarity score attributed to the analysis. At the document level, the analysis provides a list of matching sources found throughout the document for original language similarity. Each source includes a link to access it, an index number that filters the view to show only sentences flagged for that source, an Exclude button to remove it from consideration, and a similarity percentage.

Excluding matching sources at the document level will rearrange the matching sources list and alter the overall similarity percentage for original language similarity.

To help manage these matching sources, several filters are available. These filters allow users to display specific groups of matching sources, including:

• Show matching sources
• Show sources from the document
• Show sources from the internet
• Show removed sources
• Show sources with similarity below 1%

When viewing cross-institution collusion, you may be able to download the matching document from the other institution. This allows you to review the source in more detail and better understand the similarity.

The download option is only available if the other institution has enabled document sharing in its settings. If it is not enabled, the document will not be downloadable.

1.2 Sentence Level Insights (Original Language Similarity)

At the sentence level, Inspera Originality provides detailed information for every sentence flagged for original language similarity. This allows users to easily identify specific similarities and evaluate the originality of each sentence.

Each flagged sentence includes a link to the matching source and the exact sentence in the document that matches the source is also shown for context. Additionally, the system classifies the similarity type, which can be one of the following:

• Exact match
• Possibly altered text
• Cited text
• Contextual similarity

The similarity between sentences is shown as a percentage, indicating how closely they match. There is also a contextual similarity percentage, which measures how similar the surrounding context of the two sentences is. Related to this is the proximity score, which reflects the average closeness of word placement between the matched sentences.

Users have the option to change the matching source for a flagged sentence if they believe another source is more appropriate. They can also exclude a flagged sentence from the similarity analysis. When a sentence is excluded, it is unflagged and removed from the analysis, and its highlight changes to grey to show it is no longer considered.

Excluding a matching sentence will change the matching source similarity percentage as well as the overall Original Language Similarity percentage.

2. Translated Language Similarity

Translated language similarity refers to the degree of similarity detected between the document and sources in up to three other languages specified for analysis. Users can view insights for each language separately, either at the document level to see all similarity instances across the entire document in a particular language or at the sentence level to focus on specific sentences flagged for a particular language.

2.1 Document Level Insights (Translated Language Similarity)

The Translated Language Similarity Percentage represents the overall similarity between the document and sources in up to three additional languages.

Each listed language displays its own similarity percentage. Users can expand each language section to view the detected matching sources for a given language included in the analysis.

Each source includes a link to access it, an index number that filters the view to show only sentences flagged for that source, an X button to remove it from consideration, and its similarity percentage.

To help manage these matching sources, several filters are available separately for each language included in the analysis. These filters allow users to display specific groups of matching sources, including:

• Show matching sources
• Show sources from the document
• Show sources from the internet
• Show removed sources
• Show sources with similarity below 1%

Excluding a matching source for a given language will reorder the matching sources for that language, alter its similarity percentage, and impact the overall Translated Language Similarity percentage. Additionally, if the exclusion is significant, the language ranking may change, as the languages are listed based on the highest similarity percentage.

When viewing cross-institution collusion, you may be able to download the matching document from the other institution. This allows you to review the source in more detail and better understand the similarity.

The download option is only available if the other institution has enabled document sharing in its settings. If it is not enabled, the document will not be downloadable.

2.2 Sentence Level Insights (Translated Language Similarity)

At the sentence level, Inspera Originality provides detailed insights for each sentence flagged for Translated Language Similarity. This information is grouped by language, allowing users to review flagged sentences separately for each language included in the analysis.

For each language, flagged sentences include a link to the matching source, the exact matching sentence from the document, and the translated sentence from the source for context.

Additionally, the system classifies the similarity type, which can be one of the following:

• Exact match
• Possibly altered text
• Cited text
• Contextual similarity

The similarity between sentences is shown as a percentage, indicating how closely they match. There is also a contextual similarity percentage, which measures how similar the surrounding context of the two sentences is. Related to this is the proximity score, which reflects the average closeness of word placement between the matched sentences.

Users have the option to change the matching source for a flagged sentence if they believe another source is more appropriate. They can also exclude a flagged sentence for a given language included in the analysis. When a sentence is excluded, it is unflagged and removed from the analysis, and its highlight changes to grey to show it is no longer considered.

Excluding a matching sentence flagged for Translation Similarity in a specific language, the matching source percentage for that sentence will be updated. If the exclusion is significant, it can affect the ranking of matching sources in that language. The overall Translation Similarity percentage for that language will also be updated.