Small updates to L12/W12

Author: pbosilj

CMP9065 Data Programming in Python/lectures/Lecture 12 - Sklearn.ipynb

@@ -6,8 +6,6 @@
    "source": [
     "# Lecture 12 - Sklearn\n",
     "\n",
-    "# CODE 484983\n",
-    "\n",
     "[Sklearn](https://scikit-learn.org/stable/) (package name `scikit-learn`) is a machine learning library, allowing you to run a large selection of machine learning algorithms on your data.\n",
     "\n",
     "In the lectures so far (especially with Scipy), we were mostly looking at relationships between  **two features** (measurements) in a dataset  (i.e. _How is the type of vegetation related to the surface area of the lake?_). With machine learning methods, we typically want to relate **all the features** to a single **ground truth** value (this could be a class, such as an animal species, or a value to predict like the price of an item).\n",

CMP9065 Data Programming in Python/workshops/Workshop 12 - Sklearn.ipynb

@@ -105,8 +105,8 @@
     "In the following task, you are expected to train a regression model predicting the value of `petal_width` from the values of `sepal_length`. You will be training a [`linear_model.LinearRegression()`](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html) model.\n",
     "\n",
     "you can do this in the following steps:\n",
-    "- select the feature (`sepal_length`) into the variable `X_regression`. As each sample needs to be represented by an array, even if it has a single feature, you can apply `np.array.reshape(-1, 1)` to the selected feature\n",
-    "- select the feature to predict (`petal_width`) into the variable `y_regression`\n",
+    "- select the first feature (`sepal_length` - column 0) into the variable `X_regression`. As each sample needs to be represented by an array, even if it has a single feature, you can apply `np.array.reshape(-1, 1)` to the selected feature\n",
+    "- select the feature to predict (`petal_width` - column 3) into the variable `y_regression`\n",
     "- separate these into training and testing sets (`X_train, X_test, y_train, y_test`) using [`sklearn.model_selection.train_test_split()`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html), using $20\\%$ of the samples in the testing set\n",
     "- initialise the [`linear_model.LinearRegression()`](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html) model and fit it to the training data\n",
     "- predict the values for the testing set and store it into `y_predicted`\n",
@@ -209,8 +209,8 @@
     "This exercise is similar to the previous one, except that you are expected to use all other features (`sepal_length`, `sepal_width` and `petal_length`) to predict `petal_width`.\n",
     "\n",
     "You should perform the following steps:\n",
-    "- select the features (`sepal_length`, `sepal_width` and `petal_length`) into the variable `X_regression`\n",
-    "- select the feature to predict (`petal_width`) into the variable `y_regression`\n",
+    "- select the features(`sepal_length`, `sepal_width` and `petal_length` - columns 0, 1 and 2) into the variable `X_regression`\n",
+    "- select the feature to predict (`petal_width` - column 3) into the variable `y_regression`\n",
     "- separate these into training and testing sets (`X_train, X_test, y_train, y_test`) using [`sklearn.model_selection.train_test_split()`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html), using $20\\%$ of the samples in the testing set\n",
     "- initialise the [`linear_model.LinearRegression()`](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html) model and fit it to the training data\n",
     "- predict the values for the testing set and store it into `y_predicted`\n",
@@ -297,8 +297,8 @@
     "In the following task, you are expected to train a **classification model** predicting the class of the iris flower from the values of `sepal_length` and `sepal_width`. You will be training a [`tree.DecisionTreeClassifier()`](https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html) model.\n",
     "\n",
     "You can do this in the following steps:\n",
-    "- select the first two features (`sepal_length` and `sepal_width`) into the variable `X_classification`\n",
-    "- store the labels to predict `y_classification`\n",
+    "- select the first two features (`sepal_length` and `sepal_width` - column index 0 and 1) into the variable `X_classification`\n",
+    "- store the labels to predict in `y_classification` (this is currently just stored in `y`)\n",
     "- separate these into training and testing sets (`X_train, X_test, y_train, y_test`) using [`sklearn.model_selection.train_test_split()`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html), using $20\\%$ of the samples in the testing set\n",
     "- initialise the [`tree.DecisionTreeClassifier()`](https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html) model with `max_depth=5` and fit it to the training data\n",
     "- predict the values for the testing set and store it into `y_predicted`\n",
@@ -412,11 +412,12 @@
     "\n",
     "\n",
     "You can do this in the following steps:\n",
-    "- store all the dataset features into the variable `X_classification`\n",
-    "- store the labels to predict into `y_classification`\n",
+    "- store all the dataset features into the variable `X_classification` (these are currently just in `X`)\n",
+    "- store the labels to predict into `y_classification` (there are currently just in `y`)\n",
     "- separate these into training and testing sets (`X_train, X_test, y_train, y_test`) using [`sklearn.model_selection.train_test_split()`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html), using $20\\%$ of the samples in the testing set\n",
     "- initialise the [`tree.DecisionTreeClassifier()`](https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html) model with `max_depth=5` and fit it to the training data\n",
     "- predict the values for the testing set and store it into `y_predicted`\n",
+    "- **Note** that _again_ all the steps but the first (selecting the features) are **exactly the same as in [Exercise 3](#Exercise-3)**. You can copy-paste the rest of your solution for [Exercise 3](#Exercise-3) once you select the features.\n",
     "\n",
     "The code will then evaluate your model, by calculating [`accuracy_score`](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html) and [`f1_score`](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) metrics from your `y_test` and `y_predicted` classes.\n",
     "\n",
@@ -439,6 +440,12 @@
     }
    ],
    "source": [
+    "###################################\n",
+    "#### Insert your solution here ####\n",
+    "###################################\n",
+    "\n",
+    "\n",
+    "\n",
     "\n",
     "# Evaluation:\n",
     "print(\"accuracy on test set: {}\".format(accuracy_score(y_test, y_predicted)))\n",