Text Classification for DAG-Structured Categories

Due to the complex label hierarchy and intensive labeling cost in practice, the hierarchical text classification (HTC) suffers a poor performance especially when low-resource or few-shot settings are considered. Recently, there is a growing trend of applying prompts on pre-trained language models (PLMs), which has exhibited effectiveness in the few-shot flat text classification tasks. However, limited work has studied the paradigm of prompt-based learning in the HTC problem when the training data is extremely scarce. In this work, we define a path-based few-shot setting and establish a strict path-based evaluation metric to further explore few-shot HTC tasks. To address the issue, we propose the hierarchical verbalizer (“HierVerb”), a multi-verbalizer framework treating HTC as a single- or multi-label classification problem at multiple layers and learning vectors as verbalizers constrained by hierarchical structure and hierarchical contrastive learning. In this manner, HierVerb fuses label hierarchy knowledge into verbalizers and remarkably outperforms those who inject hierarchy through graph encoders, maximizing the benefits of PLMs. Extensive experiments on three popular HTC datasets under the few-shot settings demonstrate that prompt with HierVerb significantly boosts the HTC performance, meanwhile indicating an elegant way to bridge the gap between the large pre-trained model and downstream hierarchical classification tasks.

In this paper, a hierarchical ensemble classification approach that utilizes a Directed Acyclic Graph (DAG) structure is proposed as a solution to the multi-class classification problem. Two main DAG structures are considered: (i) rooted DAG, and (ii) non-rooted DAG. The main challenges that are considered in this paper are: (i) the successive misclassification issue associated with hierarchical classification, and (i) identification of the starting node within the non-rooted DAG approach. To address these issues the idea is to utilize Bayesian probability values to: select the best starting DAG node, and to dictate whether single or multiple paths should be followed within the DAG structure. The reported experimental results indicated that the proposed DAG structure is more effective than when using a simple binary tree structure for generating a hierarchical classification model.

Hierarchy-Aware and Label Balanced Model for Hierarchical Text Classification

Hierarchical multi-label text classification (HMTC) has become rather challenging when it requires handling large sets of closely related categories. We present a novel software that provides a hybrid embedding-based text representation for HMTC, shortened as HE-HMTC. It made full use of categories’ structure and their labels semantics to enrich the text representation, and therefore attempted to improve the classification performance of the text. In addition, HE-HMTC facilitates improving the accuracy of HMTC tasks. Moreover, our HE-HMTC can easily be generalized in other hierarchical classification tasks and achieve superior performance.

In this paper, we address the challenges of random label ordering and limited interpretability associated with Ensemble Classifier Chains (ECC) by introducing a novel ECC method, ECC-MOO&BN, which integrates Bayesian Networks (BN) and Multi-Objective Optimization (MOO). This approach is designed to concurrently overcome these ECC limitations. The ECC-MOO&BN method focuses on extracting diverse and interpretable label orderings for the ECC classifier. We initiated this process by employing mutual information to investigate label relationships and establish the initial structures of the BN. Subsequently, an enhanced NSGA-II algorithm was applied to develop a series of Directed Acyclic Graphs (DAGs) that effectively balance the likelihood and complexity of the BN structure. The rationale behind using the MOO method lies in its ability to optimize both complexity and likelihood simultaneously, which not only diversifies DAG generation but also helps avoid overfitting during the production of label orderings. The DAGs, once sorted topologically, yielded a series of label orderings, which were then seamlessly integrated into the ECC framework for addressing multi-label classification (MLC) problems. Experimental results show that when benchmarked against eleven leading-edge MLC algorithms, our proposed method achieves the highest average ranking across seven evaluation criteria on nine out of thirteen MLC datasets. The results of the Friedman test and Nemenyi test also indicate that the performance of the proposed method has a significant advantage compared to other algorithms.

Hierarchical text classification is an essential yet challenging subtask of multi-label text classification with a taxonomic hierarchy. Existing methods have difficulties in modeling the hierarchical label structure in a global view. Furthermore, they cannot make full use of the mutual interactions between the text feature space and the label space. In this paper, we formulate the hierarchy as a directed graph and introduce hierarchy-aware structure encoders for modeling label dependencies. Based on the hierarchy encoder, we propose a novel end-to-end hierarchy-aware global model (HiAGM) with two variants. A multi-label attention variant (HiAGM-LA) learns hierarchy-aware label embeddings through the hierarchy encoder and conducts inductive fusion of label-aware text features. A text feature propagation model (HiAGM-TP) is proposed as the deductive variant that directly feeds text features into hierarchy encoders. Compared with previous works, both HiAGM-LA and HiAGM-TP achieve significant and consistent improvements on three benchmark datasets.

Boosting ICD multi-label classification of health records with contextual embeddings and label-granularity

Text classification based on long text and multi label text classification have always been a challenge, and the combination of these two problems brings great difficulties to text classification. This study focuses on the problem of long text multi label classification. The GLM large model was used to extract abstracts from text, which effectively reduced the length of the text and retained the main content of the text. Furthermore, the Bert-BiLSTM model was used to improve the accuracy of long sequence text. This model performs particularly well in multi label classification, and can accurately classify all category labels for multi label news classification, with much higher accuracy than conventional models.

Although an improvement of hierarchical text classification can be achieved by using hierarchical structure information, existing hierarchical text classification methods suffer from two problems: data skew (especially in large-scale hierarchy) and error propagation. In this paper, we first define the concept of path-based semantic vector for the presentation of categories. Then a set of additional reliable training data for data-sparse categories can be retrieved based on such representation and particular similarity metrics. This training data enhancement strategy is classifier independent and can improve the classification of categories without adequate training data. Second, we propose the occurrence probability based strategy for hierarchical text classification which can reduce error propagation efficiently. Cooccurrence probability is then introduced to correct the errors occurred on higher levels of the hierarchy. Extensive experiments show that our hierarchical classification strategies perform well on ODP dataset, even in the condition of having few training data.

Lately, multi-label classification (MLC) problems have drawn a lot of attention in a wide range of fields including medical, web, and entertainment. The scale and the diversity of MLC problems is much larger than single-label classification problems. Especially we have to face all possible combinations of labels. To solve MLC problems more efficiently, we focus on three kinds of locality hidden in a given MLC problem. In this paper, first we show how large degree of locality exists in nine datasets, then examine how closely they are related to labels, and last propose a method of reducing the problem size using one kind of locality.

Multi-label text classification is a natural extension of text classification in which each document can be assigned with a possible widespread set of labels. Natural Language Processing (NLP) helps to understand and manipulate text in natural language by using the computer. Arabic Text Classification is challenging recently because the Arabic language is under-resourced although it has many users. The aim of this paper is to build a model to classify Arabic news and help users get and display the most relevant news to their interests. In this paper, we demonstrate the efficiency of using deep learning models in solving Arabic multi-label text classification problem. Multilayer Perceptron (MLP) and Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) have been used; we build two models via python. All data has been cleaned to improve the quality of experimental data. The result of test data in LSTM was 82.03 whereas in the MLP model was 80.37, and both models were evaluated using F1 score.

BackgroundDirected acyclic graphs (DAGs) are an effective means of presenting expert-knowledge assumptions when selecting adjustment variables in epidemiology, whereas the change-in-estimate procedure is a common statistics-based approach. As DAGs imply specific empirical relationships which can be explored by the change-in-estimate procedure, it should be possible to combine the two approaches. This paper proposes such an approach which aims to produce well-adjusted estimates for a given research question, based on plausible DAGs consistent with the data at hand, combining prior knowledge and standard regression methods.MethodsBased on the relationships laid out in a DAG, researchers can predict how a collapsible estimator (e.g. risk ratio or risk difference) for an effect of interest should change when adjusted on different variable sets. Implied and observed patterns can then be compared to detect inconsistencies and so guide adjustment-variable selection.ResultsThe proposed approach involves i. drawing up a set of plausible background-knowledge DAGs; ii. starting with one of these DAGs as a working DAG, identifying a minimal variable set, S, sufficient to control for bias on the effect of interest; iii. estimating a collapsible estimator adjusted on S, then adjusted on S plus each variable not in S in turn (“add-one pattern”) and then adjusted on the variables in S minus each of these variables in turn (“minus-one pattern”); iv. checking the observed add-one and minus-one patterns against the pattern implied by the working DAG and the other prior DAGs; v. reviewing the DAGs, if needed; and vi. presenting the initial and all final DAGs with estimates.ConclusionThis approach to adjustment-variable selection combines background-knowledge and statistics-based approaches using methods already common in epidemiology and communicates assumptions and uncertainties in a standardized graphical format. It is probably best suited to areas where there is considerable background knowledge about plausible variable relationships. Researchers may use this approach as an additional tool for selecting adjustment variables when analyzing epidemiological data.

This paper presents an investigation into the summarisation of the free text element of questionnaire data using hierarchical text classification. The process makes the assumption that text summarisation can be achieved using a classification approach whereby several class labels can be associated with documents which then constitute the summarisation. A hierarchical classification approach is suggested which offers the advantage that different levels of classification can be used and the summarisation customised according to which branch of the tree the current document is located. The approach is evaluated using free text from questionnaires used in the SAVSNET (Small Animal Veterinary Surveillance Network) project. The results demonstrate the viability of using hierarchical classification to generate free text summaries.

Hierarchical classification has been becoming a popular research topic nowadays, particularly on the web as text categorization. For a large web corpus, there can be a hierarchy with hundreds of thousands of topics, so it is common to handle this task using a flat classification approach, inducing a binary classifier only for the leaf-node classes. However, it always suffers from such low prediction accuracy due to an imbalanced issue in the training data. In this paper, we propose two novel strategies: (i) “Top-Level Pruning” to narrow down the candidate classes, and (ii) “Exclusive Top-Level Training Policy” to build more effective classifiers by utilizing the top-level data. The experiments on the Wikipedia dataset show that our system outperforms the traditional flat approach unanimously on all hierarchical classification metrics.

The Directed Acyclic Graph (DAG) is a graph representing causal pathways for informing the conduct of an observational study. The use of DAGs allows transparent communication of a causal model between researchers and can prevent over-adjustment biases when conducting causal inference, permitting greater confidence and transparency in reported causal estimates. In the era of ‘big data’ and increasing number of observational studies, the role of the DAG is becoming more important. Recent best-practice guidance for constructing a DAG with reference to the literature has been published in the ‘Evidence synthesis for constructing DAGs’ (ESC-DAG) protocol. We aimed to assess adherence to these principles for DAGs constructed within perioperative literature. Following registration on the International Prospective Register of Systematic Reviews (PROSPERO) and with adherence to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting framework for systematic reviews, we searched the Excerpta Medica dataBASE (Embase), the Medical Literature Analysis and Retrieval System Online (MEDLINE) and Cochrane databases for perioperative observational research incorporating a DAG. Nineteen studies were included in the final synthesis. No studies demonstrated any evidence of following the mapping stage of the protocol. Fifteen (79%) fulfilled over half of the translation and integration one stages of the protocol. Adherence with one stage did not guarantee fulfilment of the other. Two studies (11%) undertook the integration two stage. Unmeasured variables were handled inconsistently between studies. Only three (16%) studies included unmeasured variables within their DAG and acknowledged their implication within the main text. Overall, DAGs that were constructed for use in perioperative observational literature did not consistently adhere to best practice, potentially limiting the benefits of subsequent causal inference. Further work should focus on exploring reasons for this deviation and increasing methodological transparency around DAG construction.