Update pages/LLM_Trial_3.py

pages/LLM_Trial_3.py

@@ -1056,12 +1056,6 @@ d3=pd.DataFrame(results_df3, index=['dataset 2'])
 d4=pd.DataFrame(results_df4, index=['synthetic dataset 2'])
 df_tot= pd.concat([d1,d2,d3,d4])
 st.dataframe(df_tot)
-#st.write('These results demonstrate excellent synthetic data quality: \
-#the mean poisson deviance (mpd) and mean gamma deviance (mgd) metrics are \
-#nearly identical between real and synthetic datasets for both dataset 1 and dataset 2. \
-#This suggests the synthetic data accurately preserves the statistical properties and \
-#predictive complexity of the original data')
-
 
 # barplot comparison
 fig, ax = plt.subplots(figsize=(9, 5))
@@ -1120,9 +1114,6 @@ for container in axes[1].containers:
 
 plt.tight_layout()
 st.pyplot(fig)
-#st.write('This comparison reveals excellent synthetic data quality with minimal \
-#train-test gaps. The synthetic generation process maintains distributional properties, \
-#and also model generalization characteristics.')
 
 # Create a heatmap
 fig, ax = plt.subplots(figsize=(10, 6))
@@ -1135,8 +1126,3 @@ ax.set_ylabel('Datasets')
 
 plt.tight_layout()
 st.pyplot(fig)
-#st.write('The heatmap with the near-identical color patterns between real and synthetic versions \
-#of each dataset confirm excellent replication fidelity. Dataset 2 shows dramatically \
-#lower MPD values (green, ~0.28-0.44) compared to dataset 1 (orange-red, ~1.43-1.75), while MGD \
-#values remain similarly high across both, suggesting dataset 2 represents a different \
-#modeling challenge that the synthetic generation process successfully preserves.')