Add mean absolute percentage error (#228)

Author: Shashi456

Sources/TensorFlow/Loss.swift

@@ -77,6 +77,19 @@ public func meanAbsoluteError<Scalar: TensorFlowFloatingPoint>(
     return abs(expected - predicted).mean()
 }
 
+/// Returns the mean absolute percentage error between predictions and expectations.
+///
+/// - Parameters:
+///   - predicted: Predicted outputs from a neural network.
+///   - expected: Expected values, i.e. targets, that correspond to the correct output.
+@differentiable(wrt: predicted)
+public func meanAbsolutePercentageError<Scalar: TensorFlowFloatingPoint>(
+    predicted: Tensor<Scalar>, expected: Tensor<Scalar>
+) -> Tensor<Scalar> {
+    let diff = abs((expected - predicted) / abs(expected))
+    return 100 * diff.mean()
+}
+
 /// Returns the hinge loss between predictions and expectations.
 ///
 /// - Parameters:

Tests/TensorFlowTests/LossTests.swift

@@ -65,6 +65,15 @@ final class LossTests: XCTestCase {
         assertElementsEqual(expected: Tensor(expectedLoss), actual: loss)
     }
 
+    func testMeanAbsolutePercentageError() {
+        let predicted = Tensor<Float>([1, 2, 3, 4, 5, 6, 7, 8])
+        let expected = Tensor<Float>([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8])
+
+        let loss = meanAbsolutePercentageError(predicted: predicted, expected: expected)
+        let expectedLoss: Float = 900.0
+        assertElementsEqual(expected: Tensor(expectedLoss), actual: loss)
+    }
+
     func testMeanSquaredErrorGrad() {
         let predicted = Tensor<Float>(shape: [2, 4], scalars: [1, 2, 3, 4, 5, 6, 7, 8])
         let expected = Tensor<Float>(
@@ -232,6 +241,7 @@ final class LossTests: XCTestCase {
         ("testMeanSquaredErrorGrad", testMeanSquaredErrorGrad),
         ("testMeanSquaredLogarithmicError", testMeanSquaredLogarithmicError),
         ("testMeanAbsoluteError", testMeanAbsoluteError),
+        ("testMeanAbsolutePercentageError", testMeanAbsolutePercentageError),
         ("testHingeLoss", testHingeLoss),
         ("testKullbackLeiblerDivergence", testKullbackLeiblerDivergence),
         ("testCategoricalHingeLoss", testCategoricalHingeLoss),