Probabilistic topic models are widely used to discover latent topics in document collections, while latent feature vector representations of words have been used to obtain high performance in many NLP tasks. In this paper, we extend two different Dirichlet multinomial topic models by incorporating latent feature vector representations of words trained on very large corpora to improve the word-topic mapping learnt on a smaller corpus. Experimental results show that by using information from the external corpora, our new models produce significant improvements on topic coherence, document clustering and document classification tasks, especially on datasets with few or short documents.

Cluster analyzes have been widely used in mental health research to decompose inter-individual heterogeneity by identifying more homogeneous subgroups of individuals. However, despite advances in new algorithms and increasing popularity, there is little guidance on model choice, analytical framework and reporting requirements. In this paper, we aimed to address this gap by introducing the philosophy, design, advantages/disadvantages and implementation of major algorithms that are particularly relevant in mental health research. Extensions of basic models, such as kernel methods, deep learning, semi-supervised clustering, and clustering ensembles are subsequently introduced. How to choose algorithms to address common issues as well as methods for pre-clustering data processing, clustering evaluation and validation are then discussed. Importantly, we also provide general guidance on clustering workflow and reporting requirements. To facilitate the implementation of different algorithms, we provide information on R functions and libraries.

Knowledge distillation (KD) is a prevalent model compression technique in deep learning, aiming to leverage knowledge from a large teacher model to enhance the training of a smaller student model. It has found success in deploying compact deep models in intelligent applications like intelligent transportation, smart health, and distributed intelligence. Current knowledge distillation methods primarily fall into two categories: offline and online knowledge distillation. Offline methods involve a one-way distillation process, transferring unvaried knowledge from teacher to student, while online methods enable the simultaneous training of multiple peer students. However, existing knowledge distillation methods often face challenges where the student may not fully comprehend the teacher's knowledge due to model capacity gaps, and there might be knowledge incongruence among outputs of multiple students without teacher guidance. To address these issues, we propose a novel reciprocal teacher-student learning inspired by human teaching and examining through forward and feedback knowledge distillation (FFKD). Forward knowledge distillation operates offline, while feedback knowledge distillation follows an online scheme. The rationale is that feedback knowledge distillation enables the pre-trained teacher model to receive feedback from students, allowing the teacher to refine its teaching strategies accordingly. To achieve this, we introduce a new weighting constraint to gauge the extent of students' understanding of the teacher's knowledge, which is then utilized to enhance teaching strategies. Experimental results on five visual recognition datasets demonstrate that the proposed FFKD outperforms current state-of-the-art knowledge distillation methods.

K-nearest neighbor (KNN) rule is a simple and effective algorithm in pattern classification. In this article, we propose a local mean-based k-nearest centroid neighbor classifier that assigns to each query pattern a class label with nearest local centroid mean vector so as to improve the classification performance. The proposed scheme not only takes into account the proximity and spatial distribution of k neighbors, but also utilizes the local mean vector of k neighbors from each class in making classification decision. In the proposed classifier, a local mean vector of k nearest centroid neighbors from each class for a query pattern is well positioned to sufficiently capture the class distribution information. In order to investigate the classification behavior of the proposed classifier, we conduct extensive experiments on the real and synthetic data sets in terms of the classification error. Experimental results demonstrate that our proposed method performs significantly well, particularly in the small sample size cases, compared with the state-of-the-art KNN-based algorithms.

Early identification of pregnant women at high risk of developing gestational diabetes (GDM) is desirable as effective lifestyle interventions are available to prevent GDM and to reduce associated adverse outcomes. Personalised probability of developing GDM during pregnancy can be determined using a risk prediction model. These models extend from traditional statistics to machine learning methods; however, accuracy remains sub-optimal. We aimed to compare multiple machine learning algorithms to develop GDM risk prediction models, then to determine the optimal model for predicting GDM. A supervised machine learning predictive analysis was performed on data from routine antenatal care at a large health service network from January 2016 to June 2021. Predictor set 1 were sourced from the existing, internationally validated Monash GDM model: GDM history, body mass index, ethnicity, age, family history of diabetes, and past poor obstetric history. New models with different predictors were developed, considering statistical principles with inclusion of more robust continuous and derivative variables. A randomly selected 80% dataset was used for model development, with 20% for validation. Performance measures, including calibration and discrimination metrics, were assessed. Decision curve analysis was performed. Upon internal validation, the machine learning and logistic regression model's area under the curve (AUC) ranged from 71% to 93% across the different algorithms, with the best being the CatBoost Classifier (CBC). Based on the default cut-off point of 0.32, the performance of CBC on predictor set 4 was: Accuracy (85%), Precision (90%), Recall (78%), F1-score (84%), Sensitivity (81%), Specificity (90%), positive predictive value (92%), negative predictive value (78%), and Brier Score (0.39). In this study, machine learning approaches achieved the best predictive performance over traditional statistical methods, increasing from 75 to 93%. The CatBoost classifier method achieved the best with the model including continuous variables.