This paper proposes a convolutional deep neural network (DNN) for relation classification, which extracts lexical and sentence level features without complicated preprocessing. The method takes all word tokens as input and transforms them into vectors using word embeddings. Lexical features are extracted based on the given nouns, while sentence level features are learned using a convolutional approach. These features are concatenated to form the final feature vector, which is then fed into a softmax classifier to predict the relationship between two marked nouns. The experimental results show that the proposed method significantly outperforms state-of-the-art methods. Relation classification involves predicting semantic relations between pairs of nominals. Traditional methods rely on supervised learning and often use features derived from existing NLP systems, which can propagate errors and hinder performance. The proposed method uses a convolutional DNN to extract features directly from the input text, reducing dependency on existing NLP tools. The method includes position features (PF) to encode the relative distances to the target noun pairs, which are critical for relation classification. The system is tested on the SemEval-2010 Task 8 dataset, achieving superior performance compared to traditional methods and other DNN-based approaches. The results show that the proposed method outperforms existing techniques in terms of F1-score, indicating its effectiveness in relation classification. The method's ability to automatically learn features without relying on pre-processed data makes it a promising approach for relation classification tasks.This paper proposes a convolutional deep neural network (DNN) for relation classification, which extracts lexical and sentence level features without complicated preprocessing. The method takes all word tokens as input and transforms them into vectors using word embeddings. Lexical features are extracted based on the given nouns, while sentence level features are learned using a convolutional approach. These features are concatenated to form the final feature vector, which is then fed into a softmax classifier to predict the relationship between two marked nouns. The experimental results show that the proposed method significantly outperforms state-of-the-art methods. Relation classification involves predicting semantic relations between pairs of nominals. Traditional methods rely on supervised learning and often use features derived from existing NLP systems, which can propagate errors and hinder performance. The proposed method uses a convolutional DNN to extract features directly from the input text, reducing dependency on existing NLP tools. The method includes position features (PF) to encode the relative distances to the target noun pairs, which are critical for relation classification. The system is tested on the SemEval-2010 Task 8 dataset, achieving superior performance compared to traditional methods and other DNN-based approaches. The results show that the proposed method outperforms existing techniques in terms of F1-score, indicating its effectiveness in relation classification. The method's ability to automatically learn features without relying on pre-processed data makes it a promising approach for relation classification tasks.