Text Classification Using Document-Relational Graph Convolutional Networks

The dominant paradigm in Natural Language Processing (NLP) has traditionally relied on sequential modelling, often overlooking the rich, non-linear structural dependencies inherent in linguistic data. Graph Neural Networks (GNNs) offer a powerful alternative by modelling text as structured graphs, yet a systematic evaluation of their comparative efficacy across tasks of varying structural complexity remains under-explored. This paper presents a comprehensive comparative study of three prominent GNN architectures- Graph Convolutional Networks (GCN), Graph Attention Networks (GAT), and Graph SAGE- applied to three distinct NLP tasks: Text Classification, Semantic Role Labelling (SRL), and Abstract Meaning Representation (AMR) Parsing. We construct task-specific graph topologies, utilizing heterogeneous corpus graphs for document classification and syntactic dependency trees for SRL and AMR. Our empirical results, derived from benchmarks on the 20 Newsgroups, CoNLL-2009, and AMR 2.0 datasets, reveal a clear correlation between architectural properties and task complexity. While GCNs provide a robust and efficient baseline for global text classification (Accuracy: 96.6%), they struggle with the fine-grained structural modelling required for semantic tasks. Conversely, GATs demonstrate superior performance on syntax-heavy tasks, achieving a Labelled F1 score of 85.8% on SRL and a Smatch score of 74.5 on AMR parsing, significantly outperforming the GCN baseline (82.5% and 71.2%, respectively). These findings suggest that the anisotropic aggregation capability of attention mechanisms is critical for capturing the nuanced, long-range dependencies in natural language, establishing GAT as the preferred architecture for structure-dependent NLP applications.

Automatic building extraction from optical imagery remains a challenge due to, for example, the complexity of building shapes. Semantic segmentation is an efficient approach for this task. The latest development in deep convolutional neural networks (DCNNs) has made accurate pixel-level classification tasks possible. Yet one central issue remains: the precise delineation of boundaries. Deep architectures generally fail to produce fine-grained segmentation with accurate boundaries due to their progressive down-sampling. Hence, we introduce a generic framework to overcome the issue, integrating the graph convolutional network (GCN) and deep structured feature embedding (DSFE) into an end-to-end workflow. Furthermore, instead of using a classic graph convolutional neural network, we propose a gated graph convolutional network, which enables the refinement of weak and coarse semantic predictions to generate sharp borders and fine-grained pixel-level classification. Taking the semantic segmentation of building footprints as a practical example, we compared different feature embedding architectures and graph neural networks. Our proposed framework with the new GCN architecture outperforms state-of-the-art approaches. Although our main task in this work is building footprint extraction, the proposed method can be generally applied to other binary or multi-label segmentation tasks.

GazeGCN: Gaze-aware Graph Convolutional Network for Text Classification

The goal of text classification is to assign labels to text units accurately, which is a basic task in natural language processing. This technology has shown great value in many practical application scenarios, covering spam detection, emotional tendency analysis, user intent recognition, and many other aspects. In recent years, because of the excellent performance of graph convolutional neural networks (GCNs) in processing non-European spatial data, it has become a research hotspot in text classification and has been widely adopted. This paper first introduces the background knowledge and working principle of graph convolutional neural networks. Then the text classification method based on GCNs is described. The last thing discussed is the limitations of graph convolutional neural networks, as well as the difficulties and future development directions in this field. This paper serves as a guide for researchers and practitioners in relevant fields to fully comprehend the latest advancements in text classification methods using graph convolutional neural networks, and by revealing its limitations and challenges, it also provides guidance for future research and development work, intending to promote technical progress and application expansion in this field.

Word vector is an important tool for natural language processing (NLP) tasks such as text classification. However, existing static language models such as Word2vec cannot solve the polysemy problem, leading to a decline in text classification performance. To solve this problem, this paper proposes a method for making Chinese word vector dynamic (MCWVD). The part of speech (POS) is used to solve the ambiguity problem caused by different POS. The POS structure graph is constructed and the syntactic structure information of POS features is extracted by GCN (Graph Convolutional Network). POS vector and word vector are concatenated into PW (POS-Word) vector. Parametric matrix is added to improve the fusion effect of POS and word features. Multilayer attention is used to distinguish the importance of different features and further update the vector expression of word vectors about the current context. Experiments on Chinese datasets THUCNews and SogouNews show that MCWVD effectively improves the accuracy of text classification and achieves better performance than CoVe (Context Vectors) and ELMo (Embeddings from Language Models). MCWVD also achieves similar performance to BERT and GPT-1 (Generative Pre-Training), but with a much lower computational cost and only 4% of BERT parameters.

Text classification is a fundamental task in many applications such as topic labeling, sentiment analysis, and spam detection. The text syntactic relationship and word sequence are important and useful for text classification. How to model and incorporate them to improve performance is one key challenge. Inspired by human behavior in understanding text. In this paper, we combine the syntactic relationship, sequence structure, and semantics for text representation, and propose an attention-enhanced capsule network-based text classification model. Specifically, we use graph convolutional neural networks to encode syntactic dependency trees, build multi-head attention to encode dependencies relationship in text sequence, merge with semantic information by capsule network at last. Extensive experiments on five datasets demonstrate that our approach can effectively improve the performance of text classification compared with state-of-the-art methods. The result also shows capsule network, graph convolutional neural network, and multi-headed attention has integration effects on text classification tasks.

Aiming at the problems of diversification, complexity and islanding of power operation and inspection data and the high dependence of operation and inspection operations on expert experience and normative information, the key technology research of intelligent judgment of defect types of power operation inspection equipment is carried out. For the field of power operation and inspection, the defect text classification algorithm based on graph convolutional neural network is proposed. And the practical tests in a large defect text network diagram built by main transformer defect reports are performed. And the proposed model achieves better classification results than 7 benchmark models in the defect text classification task. Specifically, the Accuracy, Weighed-Precision, and Weighed-F1 indicators reach 73.39, 72.42, and 72.21 respectively, which improves the model’s ability to identify defect types to a greater extent and plays an important role in improving the intelligence and digitalization of power operation and inspection work.

Classification of text documents on online media is a big data problem and requires automation. Text classification accuracy can decrease if there are many ambiguous terms between classes. Hadoop Map Reduce is a parallel processing framework for big data that has been widely used for text processing on big data. The study presented text classification using genetic programming by pre-processing text using Hadoop map-reduce and collecting data using web scraping. Genetic programming is used to perform association rule mining (ARM) before text classification to analyze big data patterns. The data used are articles from science-direct with the three keywords. This study aims to perform text classification with ARM-based data pattern analysis and data collection system through web-scraping, pre-processing using map-reduce, and text classification using genetic programming. Through web scraping, data has been collected by reducing duplicates as much as 17718. Map-reduce has tokenized and stopped-word removal with 36639 terms with 5189 unique terms and 31450 common terms. Evaluation of ARM with different amounts of multi-tree data can produce more and longer rules and better support. The multi-tree also produces more specific rules and better ARM performance than a single tree. Text classification evaluation shows that a single tree produces better accuracy (0.7042) than a decision tree (0.6892), and the lowest is a multi-tree(0.6754). The evaluation also shows that the ARM results are not in line with the classification results, where a multi-tree shows the best result (0.3904) from the decision tree (0.3588), and the lowest is a single tree (0.356).

Text classification which is an integral part of text mining has caught much attention in various industries and fields recently. The ability is in assigning text documents to one or more pre-defined categories based on content similarity. While most of application of text classification focuses on document level, question classification works at much granular level such as sentence and phrase. There have been numerous studies on question classification in accordance to Bloom taxonomy in assessments to measure cognitive level of learners in higher learning institutions. But it has not been effective yet to resolve overlapping issue of Bloom taxonomy verb keywords being assigned to more than one category of Bloom taxonomy. The presence of this poses a problem in respect of classifying a particular question into a right category of Bloom taxonomy. And feature extraction plays an important role in improving the accuracy of classifier such as Support Vector Machine in question classification. Much of the past related research work focus on feature extraction methods such as bag of word (BOW) and syntactic analysis to classify questions and to address the issue, an improvement in feature extraction is needed. In view of this, this study proposes an integrated approach in feature extraction involving semantic aspect in classifying questions in accordance to Bloom taxonomy. Support Vector Machine classifier is used as it is well known for its high accuracy in text classification. With all this in place, an improved accuracy in classifying questions in accordance to Bloom taxonomy can be expected.

In the product reviews of online shopping platforms, the star rating and text comments given by the same user often appear. These data are processed by the unreasonable scoring system to give the merchants a false rating and mislead consumers. In order to improve the scoring system, an improved feature weighting method combining text review content and star rating is proposed. Firstly, the weighting rules were defined. Secondly, according to the part-of-speech evaluation function in the short text, the feature word selection was combined with the star rating and the text comment content. The CBOW model was used to train the word vector. Finally, the text classification method SVM was used to obtain the final result. This model was applied to two text datasets and compared with the traditional TF-IDF feature selection. The results show that the algorithm can effectively improve the accuracy and F1 value of text classification.

This paper proposes an ontology learning method which is used to generate a graphical ontology structure called ontology graph. The ontology graph defines the ontology and knowledge conceptualization model, and the ontology learning process defines the method of semiautomatic learning and generates ontology graphs from Chinese texts of different domains, the so-called domain ontology graph (DOG). Meanwhile, we also define two other ontological operations—document ontology graph generation and ontology graph-based text classification, which can be carried out with the generated DOG. This research focuses on Chinese text data, and furthermore, we conduct two experiments: the DOG generation and ontology graph-based text classification, with Chinese texts as the experimental data. The first experiment generates ten DOGs as the ontology graph instances to represent ten different domains of knowledge. The generated DOGs are then further used for the second experiment to provide performance evaluation. The ontology graph-based approach is able to achieve high text classification accuracy (with 92.3 % in f-measure) over other text classification approaches (such as 86.8 % in f-measure for tf–idf approach). The better performance in the comparative experiments reveals that the proposed ontology graph knowledge model, the ontology learning and generation process, and the ontological operations are feasible and effective.

Text data is one of the dominating data types in Big Data driven services and applications. The performance of text classification largely depends on the quality of feature extraction over the text corpus. For supervised learning over text documents, the TF-IDF (Term Frequency-Inverse Document Frequency) weighting factor is one of the most frequently used features in text classification. In this paper, we address two known limitations of TF-IDF based feature extraction method: First, the conventional TF-IDF weighting factor lacks of consideration about the synonymous relationship between feature terms. Second, for big corpus with large number of text documents and large number of feature terms, the computational complexity of text classification increases with the dimensionality of the feature space. We address these problems by introducing an optimization technique based on the Inter-Category Distributions (ICD) of terms and the Inter-Category Distributions of documents. We call this new weighting factor TF-IDF-ICD, namely TF-IDF with Inter-Category Distributions. To further enhance the effectiveness of our TF-IDF-ICD method, we describe a TF-IDF-ICD threshold based Dimensionality Reduction (DR) optimization. We test the text classifier with a corpus of 10, 000 articles. The evaluation results show that the proposed TF-IDF-ICD based text classification method outperforms the conventional TF-IDF based classification solution by \(7.84\%\) at only about \(43.19\%\) of the training time used by the conventional TF-IDF based text classification methods.

Analytics of machine learning-based algorithms for text classification

Short text classification task is a special kind of text classification task in that the text to be classified is generally short, typically generating a sparse text representation that lacks rich semantic information. Given this shortcoming, scholars worldwide have explored improved short text classification methods based on deep learning. However, existing methods cannot effectively use concept knowledge and long-distance word dependencies. Therefore, based on graph neural networks from the perspective of text composition, we propose concept and dependencies enhanced graph convolutional networks for short text classification. First, the co-occurrence relationship between words is obtained by sliding window, the inclusion relationship between documents and words is obtained by TF-IDF, long-distance word dependencies is obtained by Stanford CoreNLP, and the association relationship between concepts in the concept graph with entities in the text is obtained through Microsoft Concept Graph. Then, a text graph is constructed for an entire text corpus based on the four relationships. Finally, the text graph is input into graph convolutional neural networks, and the category of each document node is predicted after two layers of convolution. Experimental results demonstrate that our proposed method overall best on multiple classical English text classification datasets.

Edge features contain important information about graphs. However, current state-of-the-art neural network models designed for graph learning, e.g. graph convolutional networks (GCN) and graph attention networks (GAT), adequately utilize edge features, especially multi-dimensional edge features. In this paper, we build a new framework for a family of new graph neural network models that can more sufficiently exploit edge features, including those of undirected or multi-dimensional edges. The proposed framework can consolidate current graph neural network models; e.g. graph convolutional networks (GCN) and graph attention networks (GAT). The proposed framework and new models have the following novelties: First, we propose to use doubly stochastic normalization of graph edge features instead of the commonly used row or symmetric normalization approches used in current graph neural networks. Second, we construct new formulas for the operations in each individual layer so that they can handle multi-dimensional edge features. Third, for the proposed new framework, edge features are adaptive across network layers. As a result, our proposed new framework and new models can exploit a rich source of graph information. We apply our new models to graph node classification on several citation networks, whole graph classification, and regression on several molecular datasets. Compared with the current state-of-the-art methods, i.e. GCNs and GAT, our models obtain better performance, which testify to the importance of exploiting edge features in graph neural networks.