diff --git a/README.md b/README.md
index 1a6a7a022c1c79a7ea810d6885fc1d3483146abd..e981524b7641dcbfa87ccbcfab914561e1c4c4a7 100644
--- a/README.md
+++ b/README.md
@@ -304,3 +304,4 @@ estimators = [
 ### Conclusions
 
 I’d say we accomplished the best we could with this dataset. After trying many models with many parameters, and after trying boosting methods and stacking methods, I’d say, the we’ve reached very close to the best possible score with this dataset. Many things affect the price of the house, that we just don’t have knowledge of in this dataset, for example “vastike”, the full “huoneisto” text, more spesific location, renovations and more. The model we achieved can give a great ballpark guess of the price area, but not determine the final price without a human seller and human buyer. This could be used as online tool for people thinking of selling or buying an house, and giving them intuition on house price. And maybe give info about the model in advance, that this is more accurate on bigger cities and with flats.
+