Fix System performance section in Transparency Note

Author: psanapathi

articles/azure-health-insights/oncophenotype/transparency-note.md

@@ -71,7 +71,7 @@ Specific characteristics and limitations of the Onco Phenotype model include:
 
 ## System performance
 
-In many AI systems, performance is often defined in relation to accuracy or by  – that is, how often the AI system offers a correct prediction or output. Depending on the workflow or scenario, you can leverage the  confidence scores that are returned along with each inference and choose to set thresholds based on the tolerance for incorrect inferences. The performance of the system can be assessed by computing statistics based on true positive, true negative, false positive, and false negative instances. For example, in the tumor site predictions, one can consider a tumor site (like lung) being the positive class and other sites, including not having one, being the negative class. Using the lung tumor site as an example positive class, the following table illustrates different outcomes.
+In many AI systems, performance is often defined in relation to accuracy or by how often the AI system offers a correct prediction or output. Depending on the workflow or scenario, you can leverage the confidence scores that are returned with each inference and choose to set thresholds based on the tolerance for incorrect inferences. The performance of the system can be assessed by computing statistics based on true positive, true negative, false positive, and false negative instances. For example, in the tumor site predictions, one can consider a tumor site (like lung) being the positive class and other sites, including not having one, being the negative class. Using the lung tumor site as an example positive class, the following table illustrates different outcomes.
 
 | **Outcome**    | **Correct/Incorrect** | **Definition** | **Example** |
 | -------------- | --------------------- | -------------- | ----------- |