I like the visualization in dtreeviz because it shows the relationships between the distributions and the decision tree. The scikit-learn package has a method for displaying decision trees (see Figure 6 in Wicker and Cooper for an example), but I haven't found it to be particularly useful. Now that we have the model, we can use the dtreeviz package to create a visualization. Min_weight_fraction_leaf=0.0, presort='deprecated', Min_impurity_decrease=0.0, min_impurity_split=None, Max_depth=2, max_features=None, max_leaf_nodes=None, Time to build the model, let's instantiate a DecisionTreeClassifier.Ĭls = tree.DecisionTreeClassifier(max_depth=2)Ĭls.fit(train_X,train_y) DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini', Next, we can extract the x variables (descriptors) and the y variables (response) from the dataframe. Train_df = (train_df.clean_names(case_type="preserve") We are also taking advantage of the very cool method chaining capability provided by PyJanitor. We can use the "clean_names" function to remove the spaces from the column names. Fortunately, the PyJanitor library has functions that enable us to do all sorts of data cleaning. Note that some of the column names have spaces, which can be somewhat inconvenient. 'NumValenceElectrons', 'BalabanJ', 'BertzCT', 'Chi0', 'Chi0n', 'Chi0v', Train_df.columns Index(['Unnamed: 0', 'MolWt', 'HeavyAtomMolWt', 'NumRadicalElectrons', Let's take a look at the names of the columns in the dataframe. train_df = pd.read_csv("train_desc_with_names.csv") “not observed to crystallize” - molecules found in ZINC but not in the CSD.“observed to crystallize” - molecules that occur in both ZINC and the Cambridge Crystallographic Database (CSD).In their paper, Wicker and Cooper use a set of 40,541 commercially available molecules, from the ZINC database, to establish a relationship between molecular flexibility and the ability of a molecule to crystallize. This post will also show off a couple of useful Python libraries that I've recently integrated into my workflow.
DECISION TREE VISUALIZATION PYTHON HOW TO
In this spirit, I thought I'd put together a quick post showing how to build and visualize a decision tree. They can also provide a means of understanding the relationship between sets of experiments, particularly with pharmacokinetic data. Decision trees can often provide an efficient way of looking at the relationship between molecular descriptors and experimental data. This paper reminded me of the power of a simple decision tree. A 2016 paper by Wicker and Cooper, describing a molecular descriptor designed to capture molecular flexibility, popped up on Twitter this week.
0 kommentar(er)
