Update app.py

app.py

@@ -10,7 +10,7 @@ from annotator import ignitor_load,appendor,allll
 
 if __name__ == '__main__':
     st.header("Which sentiment does this text convey? Is it related to work?")
-    st.write("This app tells whether a text can be labeled as related to pecuniary, colleagues, boss, temporal/general aspects of work, or neither of them. It also shows the sentiment of text across 27 different affects + neutral. The goal of this app is to be served as a proof of concept and cross-validator ( data validator ) of the tweets classified by the full-fledged model. The purpose of this model is to recognize tweets which are related to work from within large corpora of data, to see how their sentiment differs from the rest of the tweets. The FN of the model is high to prevent labeling tweets that are not related to work, as long as the sentiment of labeled tweets is not systematically different than the FN tweets, the goal of this model is met. To mitigate the risk of systematic wrong-classification, in the work-classification phase, none of the keywords conveys a sentiment.") 
+    st.write("This app tells whether a text can be labeled as related to pecuniary, colleagues, boss, temporal/general aspects of work, or neither of them. It also shows the sentiment of text across 27 different affects + neutral. The goal of this app is to be served as a proof of concept and cross-validator ( data validator ) of the tweets classified by the full-fledged model and to show the code and method that we created and used in our paper. The purpose of this model is to recognize tweets which are related to work from within large corpora of data, to see how their sentiment differs from the rest of the tweets. The FN of the model is high to prevent labeling tweets that are not related to work, as long as the sentiment of labeled tweets is not systematically different than the FN tweets, the goal of this model is met. To mitigate the risk of systematic wrong-classification, in the work-classification phase, none of the keywords conveys a sentiment.") 
     st.write("You should expect that around 4 percent of all the tweets **containing potentially work-related words**, to be classified as work-related by this app.")
     st.write("This model is powered by Joe Davison GoEmotions-based model and HuggingFace for the sentiment-labeling phase and for the work-labeling phase it is powered by HuggingFace, DistilBERT, Streamlit and, the Active Label Selection (ALS) method developed by Sepehr Cyrusian as a part of the article co-written by professor Colin P. Green")