diff --git a/src/shogun/clustering/KMeansLloydImpl.h b/src/shogun/clustering/KMeansLloydImpl.h
index 9c411752eeb..9ebd294fbc1 100644
--- a/src/shogun/clustering/KMeansLloydImpl.h
+++ b/src/shogun/clustering/KMeansLloydImpl.h
@@ -29,8 +29,6 @@ class CKMeansLloydImpl
 		 * @param distance distance
 		 * @param max_iter max iterations allowed
 		 * @param mus cluster centers matrix (k columns)
-		 * @param ClList cluster number each data vector belongs (size no_of_vectors)
-		 * @param weights_set no. of points belonging to each cluster (size k)
 		 * @param fixed_centers keep centers fixed or not
 		 */
 		static void Lloyd_KMeans(int32_t k, CDistance* distance, int32_t max_iter, SGMatrix<float64_t> mus,
diff --git a/src/shogun/lib/external/cdflib.cpp b/src/shogun/lib/external/cdflib.cpp
index 1d2c8cb95da..ccc4f770b76 100644
--- a/src/shogun/lib/external/cdflib.cpp
+++ b/src/shogun/lib/external/cdflib.cpp
@@ -10861,11 +10861,9 @@ void timestamp ( )
 # define TIME_SIZE 40
 
   static char time_buffer[TIME_SIZE];
-  const struct tm *tm;
   time_t now;
 
   now = time ( NULL );
-  tm = localtime ( &now );
 
 
   cout << time_buffer << "\n";
diff --git a/src/shogun/machine/LinearMachine.h b/src/shogun/machine/LinearMachine.h
index ec477158fd6..9b9dd041dd2 100644
--- a/src/shogun/machine/LinearMachine.h
+++ b/src/shogun/machine/LinearMachine.h
@@ -192,7 +192,7 @@ class CLinearMachine : public CMachine
 		float64_t bias;
 		/** features */
 		CDotFeatures* features;
-		/** If true, bias is computed in ::train method */
+		/** If true, bias is computed in train method */
 		bool m_compute_bias;
 };
 }
diff --git a/src/shogun/mathematics/linalg/internal/implementation/ElementwiseSquare.h b/src/shogun/mathematics/linalg/internal/implementation/ElementwiseSquare.h
index 2731ccb0442..c842e4d4ca5 100644
--- a/src/shogun/mathematics/linalg/internal/implementation/ElementwiseSquare.h
+++ b/src/shogun/mathematics/linalg/internal/implementation/ElementwiseSquare.h
@@ -128,7 +128,7 @@ struct elementwise_square<Backend::EIGEN3,Matrix>
 	/**
 	 * Method that computes the square of co-efficients of a dense matrix
 	 *
-	 * @param m the matrix whose squared co-efficients matrix has to be computed
+	 * @param mat the matrix whose squared co-efficients matrix has to be computed
 	 * @param result Pre-allocated matrix for the result of the computation
 	 */
 	static void compute(SGMatrix<T> mat, SGMatrix<T> result)
diff --git a/src/shogun/mathematics/linalg/internal/implementation/MeanEigen3.h b/src/shogun/mathematics/linalg/internal/implementation/MeanEigen3.h
index ee28a9118f3..9eb8f6b03ed 100644
--- a/src/shogun/mathematics/linalg/internal/implementation/MeanEigen3.h
+++ b/src/shogun/mathematics/linalg/internal/implementation/MeanEigen3.h
@@ -153,7 +153,7 @@ struct mean<Backend::EIGEN3, Matrix>
 	/**
 	 * Method that computes the mean of SGVectors using Eigen3
 	 *
-	 * @param a vector whose mean has to be computed
+	 * @param vec a vector whose mean has to be computed
 	 * @return the vector mean \f$\bar a_i\f$
 	 */
 	static ReturnType compute(SGVector<T> vec)
diff --git a/src/shogun/neuralnets/RBM.h b/src/shogun/neuralnets/RBM.h
index 438a65be85f..9b7d1116970 100644
--- a/src/shogun/neuralnets/RBM.h
+++ b/src/shogun/neuralnets/RBM.h
@@ -301,7 +301,7 @@ friend class CDeepBeliefNetwork;
 	 * during computation. If not given, a new matrix is allocated and used as
 	 * a buffer.
 	 *
-	 * @param return Approximation to the average pseudo-likelihood over the
+	 * @return Approximation to the average pseudo-likelihood over the
 	 * given batch
 	 */
 	virtual float64_t pseudo_likelihood(SGMatrix<float64_t> visible,
diff --git a/src/shogun/statistics/NOCCO.h b/src/shogun/statistics/NOCCO.h
index fc40f7f9995..7989a95ea55 100644
--- a/src/shogun/statistics/NOCCO.h
+++ b/src/shogun/statistics/NOCCO.h
@@ -96,7 +96,7 @@ template<class T> class SGMatrix;
  *	\mathbf R_X^i\cdot \mathbf R_Y^i
  * \f]
  *
- * For performing the independence test, ::PERMUTATION test is used by first
+ * For performing the independence test, PERMUTATION test is used by first
  * randomly shuffling the samples from one of the distributions while keeping
  * the samples from the other distribution in the original order. This way we
  * sample the null distribution and compute p-value and threshold for a given