# `Latu.ML.Classification.RandomForestClassificationModel`
[🔗](https://github.com/zero-one-group/latu_ml/blob/v0.4.2/lib/latu/ml/classification.ex#L0)

Attributes of a fitted `RandomForestClassificationModel`.

Every accessor here is one name on the server's allowlist for
`org.apache.spark.ml.classification.RandomForestClassificationModel` — 14 of them — so
tab completion is that allowlist. The server refuses anything else with
`CONNECT_ML.ATTRIBUTE_NOT_ALLOWED`; these refuse a model of the wrong class before it gets
that far, and name the module that would have taken it.

Allowed by the server but not here: `toString` is `Identifiable`'s, and tells you less than
the model's own uid; `evaluate` answers with a training summary, which has no struct yet.
`Latu.ML.attribute/2` and `Latu.ML.attribute/3` will send any of these names; it is what
comes back that has nowhere to go.

The allowlist is inherited rather than per class: this one is the union of
`org.apache.spark.ml.PredictionModel`,
`org.apache.spark.ml.classification.ClassificationModel`,
`org.apache.spark.ml.classification.ProbabilisticClassificationModel`,
`org.apache.spark.ml.classification.RandomForestClassificationModel`,
`org.apache.spark.ml.tree.TreeEnsembleModel`, `org.apache.spark.ml.util.HasTrainingSummary`,
`org.apache.spark.ml.util.Identifiable`.

# `class`

```elixir
@spec class() :: String.t()
```

The JVM class these accessors belong to.

# `feature_importances`

```elixir
@spec feature_importances(Latu.ML.Model.t()) ::
  {:ok, term()} | {:error, Latu.Error.t()}
```

Estimate of the importance of each feature.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `get_num_trees`

```elixir
@spec get_num_trees(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Number of trees in ensemble.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `has_summary`

```elixir
@spec has_summary(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Indicates whether a training summary exists for this model instance.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `num_classes`

```elixir
@spec num_classes(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Number of classes (values which the label can take).

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `num_features`

```elixir
@spec num_features(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Returns the number of features the model was trained on. If unknown, returns -1.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `predict`

```elixir
@spec predict(Latu.ML.Model.t(), Nx.Tensor.t() | Latu.ML.SparseVector.t()) ::
  {:ok, term()} | {:error, Latu.Error.t()}
```

Predict label for the given features.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

## Arguments

  * `value` — a Spark `Vector`: an `Nx.Tensor` or a `Latu.ML.SparseVector`

# `predict_leaf`

```elixir
@spec predict_leaf(Latu.ML.Model.t(), Nx.Tensor.t() | Latu.ML.SparseVector.t()) ::
  {:ok, term()} | {:error, Latu.Error.t()}
```

Predict the indices of the leaves corresponding to the feature vector.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

## Arguments

  * `value` — a Spark `Vector`: an `Nx.Tensor` or a `Latu.ML.SparseVector`

# `predict_probability`

```elixir
@spec predict_probability(Latu.ML.Model.t(), Nx.Tensor.t() | Latu.ML.SparseVector.t()) ::
  {:ok, term()} | {:error, Latu.Error.t()}
```

Predict the probability of each class given the features.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

## Arguments

  * `value` — a Spark `Vector`: an `Nx.Tensor` or a `Latu.ML.SparseVector`

# `predict_raw`

```elixir
@spec predict_raw(Latu.ML.Model.t(), Nx.Tensor.t() | Latu.ML.SparseVector.t()) ::
  {:ok, term()} | {:error, Latu.Error.t()}
```

Raw prediction for each possible label.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

## Arguments

  * `value` — a Spark `Vector`: an `Nx.Tensor` or a `Latu.ML.SparseVector`

# `summary`

```elixir
@spec summary(Latu.ML.Model.t()) :: Latu.ML.Summary.t()
```

Gets summary of the model trained on the training set. An exception is thrown if no summary
exists.

A lazy builder: Spark reaches a summary through the model that owns it, so the reference is
composed here and nothing is sent. `Latu.ML.summary/1` is the same thing without the class
check.

# `to_debug_string`

```elixir
@spec to_debug_string(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Full description of model.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `total_num_nodes`

```elixir
@spec total_num_nodes(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Total number of nodes, summed over all trees in the ensemble.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `tree_weights`

```elixir
@spec tree_weights(Latu.ML.Model.t()) :: {:ok, term()} | {:error, Latu.Error.t()}
```

Return the weights for each tree.

An action: it reaches the server. A `Vector` or `Matrix` comes back as an `Nx.Tensor`, or a
`Latu.ML.SparseVector` where densifying would be this package's decision rather than yours.

# `trees`

```elixir
@spec trees(Latu.ML.Model.t()) ::
  {:ok, [Latu.ML.Model.t()]} | {:error, Latu.Error.t()}
```

Trees in this ensemble. Warning: These have null parent Estimators.

An action, and it **spends cache**: the server registers every sub-model on its way out, so
this answers with a list of `Latu.ML.Model` — one `DecisionTreeClassificationModel` and
one cache entry per tree — and they are yours to give back. `Latu.ML.delete/1` takes the
whole list in one call.

---

*Consult [api-reference.md](api-reference.md) for complete listing*