Example of classification

docs/getting-started/first-program.md

@@ -1,3 +1,129 @@
 # Your First Safe-DS Program
-test
-TODO
+
+## Classification
+[Link to Titanic Dataset](https://github.com/Safe-DS/Datasets/blob/main/src/safeds_datasets/tabular/_titanic/data/titanic.csv)
+
+
+The Titanic dataset is a simple example for your first machine learning project. The dataset contains data about passengers who traveled on the Titanic. Obviously, not all passengers had the same chances of survival. This model is intended to generalize, based on the data, what characteristics characterize a person who has survived.
+
+
+```sds
+Package Titanic
+
+Pipeline Titanic{
+    // load the table from CSV file 
+    val data = Table.fromCsvFile("titanic.csv"); 
+
+    // using a mean imputer to fill up NULL-values with the Mean
+    val meanImputer = SimpleImputer(SimpleImputer.Strategy.Mean); 
+    val impMean, val titanicMeanImputed = meanImputer.fitAndTransform(titanicRemoved, ["age", "fare"]);
+
+    //using a Median imputer to fill up NULL-values with the Median
+    val medianImputer  = SimpleImputer(SimpleImputer.Strategy.Mode);
+    val impMedian, val titanicMedianImputed = medianImputer.fitAndTransform(titanicMeanImputed, ["port_embarked"]); 
+
+    /*Using a onehotencoder to convert categorical data into a binary
+        format allows each category to be represented as a unique combination
+        of binary variables, making it easier for machine learning algorithms 
+        to process and interpret the data.*/
+    val oneHotEncoder = OneHotEncoder(); 
+    val enc, val titanicEncoded = oneHotEncoder.fitAndTransform(titanicMedianImputed, ["sex", "port_embarked"]);
+
+    //using a discretizer to bin continuous data into intervals.
+    val discretisizer = Discretizer();
+    val disc, val titanicDiscretisized = discretisizer.fitAndTransform(titanicEncoded, ["fare"]);
+
+    //using the datastructure plot to generate a heatmap
+	val plot = titanic.plot.correlationHeatmap();
+
+    /*Splitting the data and tagging the Target: 70% of the data 
+        was used for training. The remaining 30% of the data was used for 
+        testing.*/
+    val trainigSet, val testSet = titanicDiscretisized.splitRows(0.7);
+    val trainingTable = trainigSet.toTabularDataset("survived");
+	val testTable = testSet.toTabularDataset("survived");
+
+    //training the model with a Gradient Boosting Classifier
+    val gbClassifier = GradientBoostingClassifier(10,0.2);
+    val fittedgb = gbClassifier.fit(trainingTable);
+    //calculating the accuracy
+    val accuracygb = gbFitted.accuracy(testTable);
+}
+```
+###[Table](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/containers/Table/#safeds.data.tabular.containers.Table) & [TabularDataset](https://dsl.safeds.com/en/stable/api/safeds/data/labeled/containers/TabularDataset/#safeds.data.labeled.containers.TabularDataset)
+
+
+A Table is a two-dimensional collection of data. It can either be seen as a list of rows or as a list of columns.
+
+A TabularDataSet is a dataset containing tabular data. It can be used to train machine learning models.
+
+
+```sds
+    val data = Table.fromCsvFile("titanic.csv"); 
+    //...
+    val trainigSet, val testSet = titanicDiscretisized.splitRows(0.7);
+    val trainingTable = trainigSet.toTabularDataset("survived");
+	val testTable = testSet.toTabularDataset("survived");
+```
+
+###[SimpleImputer](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/transformation/SimpleImputer/#safeds.data.tabular.transformation.SimpleImputer)
+
+
+Replace missing values with the given strategy.
+
+
+```sds
+    val meanImputer = SimpleImputer(SimpleImputer.Strategy.Mean); 
+    val impMean, val titanicMeanImputed = meanImputer.fitAndTransform(titanicRemoved, ["age", "fare"]);
+
+    val medianImputer  = SimpleImputer(SimpleImputer.Strategy.Mode);
+    val impMedian, val titanicMedianImputed = medianImputer.fitAndTransform(titanicMeanImputed, ["port_embarked"]); 
+```
+
+###[OneHotEncoder](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/transformation/OneHotEncoder/#safeds.data.tabular.transformation.OneHotEncoder)
+
+
+A way to deal with categorical features that is particularly useful for unordered (i.e. nominal) data.
+
+
+```sds
+    val oneHotEncoder = OneHotEncoder(); 
+    val enc, val titanicEncoded = oneHotEncoder.fitAndTransform(titanicMedianImputed, ["sex", "port_embarked"]);
+```
+
+###[Discretizer](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/transformation/Discretizer/#safeds.data.tabular.transformation.Discretizer)
+
+
+The Discretizer bins continuous data into intervals.
+
+
+```sds
+    val discretisizer = Discretizer();
+    val disc, val titanicDiscretisized = discretisizer.fitAndTransform(titanicEncoded, ["fare"]);
+```
+
+###[Plots](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/plotting/TablePlotter/#safeds.data.tabular.plotting.TablePlotter)
+We used the [CorrelationHeatMap](https://dsl.safeds.com/en/stable/api/safeds/data/tabular/plotting/TablePlotter/#safeds.data.tabular.plotting.TablePlotter.correlationHeatmap)
+
+
+A class that contains plotting methods for a table.
+
+
+```sds
+	val plot = titanic.plot.correlationHeatmap();
+```
+
+
+
+### Classifier
+
+A classifier is a machine learning algorithm that categorizes data into predefined classes. It is trained on labeled data to learn patterns and then used to predict the class of new, unseen data.
+
+
+In our excample we used the [GradientBoostingClassifier](http://127.0.0.1:8000/api/safeds/ml/classical/classification/GradientBoostingClassifier/#safeds.ml.classical.classification.GradientBoostingClassifier)
+
+```sds
+    val gbClassifier = GradientBoostingClassifier(10,0.2);
+    val fittedgb = gbClassifier.fit(trainingTable);
+    val accuracygb = gbFitted.accuracy(testTable);
+```