The Operator provides different filter types to make Attribute selection easy. Possibilities are for example: Direct selection of Attributes. Selection by a regular expression or selecting only Attributes without missing values. See parameter attribute filter type for a detailed description of the different filter types. Show
The type parameter can be used to decide whether to include or exclude the selected Attributes. Special Attributes (Attributes with Roles, like id, label, weight) are by default ignored in the selection. They will always remain in the resulting output ExampleSet. The parameter also apply to special attributes changes this. Only the selected Attributes are delivered to the output port. The rest are removed from the ExampleSet. DifferentiationSelect by <...> Operators There are several Operators that select Attributes according to their input. For example Select by Weights selects Attributes whose weights match a specified criterion. The Select by Random Operator selects a random subset of Attributes. Remove Attribute Range removes a range of Attributes according to the index of the Attributes. The Remove Useless Attributes Operator removes Attributes which can be considered to be useless according to some specified criteria. The Remove Correlated Attributes Operator removes Attributes which are correlated to each other. Work on SubsetThis Operator is a combination of the Select Attributes Operator and the Subprocess Operator. It applies the Operators in its inner process to an ExampleSet with only the Attributes which are selected by the attribute filter type. The inner result is merged back to the whole input ExampleSet. Forward SelectionThis is an implementation of the forward selection feature selection method. It selects the most relevant Attributes according to a model which is trained inside the Operator. For details see the documentation of the Forward Selection Operator. Backward EliminationThis is an implementation of the backward elimination feature selection method. It selects the most relevant Attributes according to an model which is trained inside the Operator. For details see the documentation of the Forward Selection Operator. Filter ExamplesThis Operator does not select Attributes, but filters (or select) Examples. Thus, it does what Select Attributes does but applied to Examples instead of Attributes. Input
Output
Parameters
Tutorial ProcessesSelecting Attributes from the Titanic Data SampleThis tutorial Process shows the basic usage of the Select Attributes Operator. First the 'Titanic' data is retrieved from the Samples folder. The first Select Attributes Operator selects a subset of the Attributes. The subset is specified by the select subset parameter. The original output port is connected to the input port of the second Select Attributes Operator. There, only nominal Attributes are selected by choosing binominal and non-binominal. Different usages of the Select Attributes OperatorThis tutorial Process demonstrates different usages of the Select Attributes Operator. A demo ExampleSet is created inside a Subprocess Operator. It has 3 special Attributes (id, label, weight) and 5 regular Attributes (att1, att2, att3, att4, att5). Also different attribute types are used (integer: id; binominal: label; real: weight, att1, att2, att4, att5; nominal: att3). After the Subprocess Operator a Breakpoint is inserted, to investigate the demo ExampleSet. Next several Select Attributes Operators are used to show the different attribute filter types and the combinations with the parameters type and also apply to special attributes. See the comments in the process for more details. Selecting Attributes by using a regular expressionThis tutorial Process illustrates the usage of a regular expression to select Attributes from the Labor-Negotiations data sample. The regular expression specified is: w.*|.*y.* This means all Attributes starting with a 'w' (w.*) or (|) all Attributes whose names contain a 'y' in their name (.*y.*) match the expression. The following Attributes of the Labor-Negotiations data set match this expression: wage-inc-1st, wage-inc-2nd, wage-inc-3rd, working-hours, standby-pay, statutory-holidays, longterm-disability-assistance. The Attributes that match the condition in the exclude expression parameter will be removed. The specified exclude expression is: .*\[0-9\].*. This means all Attributes whose name contains a digit are removed. Finally the following four Attributes are selected: working-hours, standby-pay, statutory-holidays, longterm-disability-assistance. Beside these, the special Attribute class is also kept. Apa itu atribut pada Python?Atribut adalah data anggota (variabel kelas dan variabel contoh) dan metode, diakses melalui notasi titik. Sebuah variabel yang dibagi oleh semua contoh kelas. Variabel kelas didefinisikan dalam kelas tapi di luar metode kelas manapun.
Apa fungsi class pada Python?Class atau kelas-kelas adalah prototipe untuk wadah menghimpun data dan kebergunaan yang kemudian menghasilkan objek. Setiap class baru akan menghasilkan objek baru yang kemudian bisa dibuat instance dengan memiliki atribut yang ada.
Apa yang dimaksud constructor pada class Python?Apa itu konstruktor? Konstruktor adalah sebuah fungsi yang akan dipanggil pertama kali saat sebuah objek di-instantiasi-kan. Fungsi tersebut harus selalu bernama __init__() .
Apa yang dimaksud dengan namespace pada Python?Sebuah namespace adalah pemetaan dari nama ke objek. Sebagian besar ruang nama namespace saat ini diimplementasikan sebagai kamus dictionary Python, tetapi itu biasanya tidak terlihat dengan cara apa pun (kecuali untuk kinerja), dan itu mungkin berubah di masa depan.
|