update doc

Author: cokelaer

doc/data.rst

@@ -37,11 +37,14 @@ Quote characters are also removed.
 IC50
 ------
 
-The specific format **IC50** is CSV file where the header must 
+The specific format for the **IC50** data is CSV file where the header must 
 contain a column named **COSMIC_ID**. Other
 column should correspond to a given drug identifier (an integer).  The order
 of the columns does not matter. So, each row contains the IC50s for a given COSMIC identifier.
 
+.. warning:: the data read in the CSV file is not transformed. So, IC50 data should be 
+   logged data.
+
 
 .. note:: The IC50 matrix can be populated with other data (e.g., AUCs).
 

doc/regression.rst

@@ -51,9 +51,9 @@ drug, we tune the **alpha** parameter using a cross validation (a K-fold CV).
 
 
 .. note:: the regression implementation is based on scikit-learn. The **alpha**
-   parameter is called **lambda** in the R glmnnet package. In the ElasticNet
+   parameter is called **lambda** in the R glmnet package. In the ElasticNet
    case, and additional parameter is called **l1_ratio** (0.5 by default)
-   and correspond to the  glmnet parameter **alpha**.
+   and corresponds to the  glmnet parameter **alpha**.
 
 Before, doing the tuning, let us choose one drug. Let us pick up an interesting
 one::