Implementation of a machine learning algorithm for automated thematic annotations in avatar: A linear support vector classifier approach

Hudon, Alexandre; Beaudoin, Mélissa; Phraxayavong, Kingsada; Dellazizzo, Laura; Potvin, Stéphane; Dumais, Alexandre

doi:10.1177/14604582221142442

Cited by 11 publications

(16 citation statements)

References 38 publications

(54 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance for these initial automated annotations for the Avatar and the patients ‘themes is presented in Table 3 . The results are comparable to the ones obtained in the description of the original study on automated classification of interaction for AT [ 33 ].…”

Section: Methodssupporting

confidence: 84%

See 1 more Smart Citation

Dyadic Interactions of Treatment-Resistant Schizophrenia Patients Having Followed Virtual Reality Therapy: A Content Analysis

Hudon

Couture

Dellazizzo

et al. 2023

JCM

Self Cite

View full text Add to dashboard Cite

(1) Background: Very little is known about the inner therapeutic processes of psychotherapy interventions for patients suffering from treatment-resistant schizophrenia. Avatar therapy (AT) is one such modalities in which the patient is undergoing immersive sessions in which they interact with an Avatar representing their main persistent auditory verbal hallucination. The aim of this study is to identify the most prevalent dyadic interactions between the patient and the Avatar in AT for patient’s suffering from TRS. (2) Methods: A content analysis of 256 verbatims originating from 32 patients who completed AT between 2017 and 2022 at the Institut universitaire en santé mentale de Montréal was conducted to identify dyadic interactions between the patients and their Avatar. (3) Results: Five key dyads were identified to occur on average more than 10 times for each participant during the immersive sessions across their AT: (Avatar: Reinforcement, Patient: Self-affirmation), (Avatar: Provocation, Patient: Self-affirmation), (Avatar: Coping mechanisms, Patient: Prevention), (Patient: Self-affirmation, Avatar: Reinforcement), and (Patient: Self-appraisal, Avatar: Reinforcement). (4) Conclusion: These dyads offer a first qualitative insight to the interpersonal dynamics and patient-avatar relationships taking place during AT. Future studies on the implication of such dyadic interactions with the therapeutic outcome of AT should be conducted considering the importance of dyadic relationships in psychotherapy.

show abstract

Section: Methodssupporting

confidence: 84%

“…The annotation of the verbatims was done automatically by using a peer-reviewed trained linear support vector classifier, previously trained and implemented on a dataset for AT using Python 3.9 with the Scitkit-Learn open library and a 10-k fold cross validation [ 33 ].…”

Section: Methodsmentioning

confidence: 99%

Dyadic Interactions of Treatment-Resistant Schizophrenia Patients Having Followed Virtual Reality Therapy: A Content Analysis

Hudon

Couture

Dellazizzo

et al. 2023

JCM

Self Cite

View full text Add to dashboard Cite

show abstract

“…Development of Various Other Machine Learning c-RASAR Models. The combination of RASAR descriptors obtained after the most discriminating feature analysis was used for developing additional ML models applying various algorithms like logistic regression (LR), 32 linear support vector classifier (LSVC), 33 support vector classifier (SVC), 34 decision tree (DT), 35 random forest classifier (RFC), 36 knearest neighbor classifier (k-NN), 37 and Gaussian Nai ̈ve Bayes (NB) 38 using the basic setting of the associated hyperparameters to compare the predictive ability among the different ML classifiers. Some of these classification modeling tools are also available from the DTC Lab Web site page https://sites.google.com/jadavpuruniversity.in/dtc-labsoftware/home/machine-learning-model-development-guis.…”

Section: ■ Materials and Methodsmentioning

confidence: 99%

Prediction-Inspired Intelligent Training for the Development of Classification Read-across Structure–Activity Relationship (c-RASAR) Models for Organic Skin Sensitizers: Assessment of Classification Error Rate from Novel Similarity Coefficients

Banerjee

Roy

2023

Chem. Res. Toxicol.

View full text Add to dashboard Cite

The advancements in the field of cheminformatics have led to a reduction in animal testing to estimate the activity, property, and toxicity of query chemicals. Read-across structure− activity relationship (RASAR) is an emerging concept that utilizes various similarity functions derived from chemical information to develop highly predictive models. Unlike quantitative structure− activity relationship (QSAR) models, RASAR descriptors of a query compound are computed from its close congeners instead of the compound itself, thus targeting predictions in the model training phase. The objective of the present study is not to propose new QSAR models for skin sensitization but to demonstrate the enhancement in the quality of predictions of the skin-sensitizing potential of organic compounds by developing classification-based RASAR (c-RASAR) models. A diverse, previously curated data set was collected from the literature for which 2D descriptors were computed. The extracted essential features were then used to develop a classification-based linear discriminant analysis (LDA) QSAR model. Furthermore, from the read-across-based predictions, RASAR descriptors were calculated using the basic settings of the hyperparameters for the Laplacian Kernel-based optimum similarity measure. After feature selection, an LDA c-RASAR model was developed, which superseded the prediction quality of the LDA−QSAR model. Various other combinations of RASAR descriptors were also taken to develop additional c-RASAR models, all showing better prediction quality than the LDA QSAR model while using a lower number of descriptors. Various other machine learning c-RASAR models were also developed for comparison purposes. In this work, we have proposed and analyzed three new similarity metrics: g m _class, s m 1 , and s m 2 . The first one is an indicator variable used to generate a simple univariate c-RASAR model with good prediction ability, while the remaining two are similarity indices used to analyze possible activity cliffs in the training and test sets and are believed to play an important role in the modelability analysis of data sets.

show abstract

“…The combination of RASAR descriptors obtained after the most discriminating feature analysis was used for developing additional ML models applying various algorithms like Logistic Regression (LR) [29], Linear Support Vector Classifier (LSVC) [30], Support Vector Classifier (SVC) [31], Decision Tree (DT) [32], Random Forest Classifier (RFC) [33], k-Nearest Neighbor classifier (k-NN) [34] and Gaussian Naïve Bayes (NB) [35] using the basic setting of the associated hyperparameters to compare the predictive ability among the different ML classifiers. Some of these classification modeling tools are also available from the DTC Lab website page https://sites.google.com/jadavpuruniversity.in/dtc-lab-software/home/machine-learning-modeldevelopment-guis.…”

Section: Development Of Various Other Machine Learning (Ml) C-rasar M...mentioning

confidence: 99%

Prediction-inspired intelligent training for the development of c-RASAR models for organic skin sensitizers: Assessment of classification error rate from novel similarity coefficients

Banerjee

Roy

2023

Preprint

View full text Add to dashboard Cite

The advancements in the field of cheminformatics have led to a reduction in animal testing to estimate the activity/property/toxicity of query chemicals. Read-Across Structure-Activity Relationship (RASAR) is an emerging concept that utilizes various similarity functions derived from chemical information to develop highly predictive models. Unlike quantitative structure-activity relationship (QSAR) models, RASAR descriptors of a query compound are computed from its close congeners instead of the compound itself, thus targeting predictions in the model training phase. The objective of the present study is not to propose new QSAR models for skin sensitization, but to demonstrate the enhancement in the quality of predictions of the skin-sensitizing potential of organic compounds by developing classification-based RASAR (c-RASAR) models. A diverse, previously curated, dataset was collected from the literature, for which 2D descriptors were computed. The extracted essential features were then used to develop a classification-based linear discriminant analysis (LDA) QSAR model. Furthermore, from the Read-Across-based predictions, RASAR descriptors were calculated using the basic settings of the hyperparameters for the Laplacian kernel-based optimum similarity measure. After feature selection, an LDA c-RASAR model was developed which superseded the prediction quality of the LDA-QSAR model. Various other combinations of RASAR descriptors were also taken to develop additional c-RASAR models all showing better prediction quality than the LDA QSAR model while using a lower number of descriptors. Various other machine learning c-RASAR models were also developed for comparison purposes. In this work, we have proposed and analyzed three new similarity metrics: gm_class, sm1, and sm2. The first one is an indicator variable used to generate a simple univariate c-RASAR model with good prediction ability, while the rest two are similarity indices used to analyze possible activity cliffs in the training and test sets and are believed to play an important role in the modelability analysis of datasets.

show abstract

Implementation of a machine learning algorithm for automated thematic annotations in avatar: A linear support vector classifier approach

Cited by 11 publications

References 38 publications

Dyadic Interactions of Treatment-Resistant Schizophrenia Patients Having Followed Virtual Reality Therapy: A Content Analysis

Dyadic Interactions of Treatment-Resistant Schizophrenia Patients Having Followed Virtual Reality Therapy: A Content Analysis

Prediction-Inspired Intelligent Training for the Development of Classification Read-across Structure–Activity Relationship (c-RASAR) Models for Organic Skin Sensitizers: Assessment of Classification Error Rate from Novel Similarity Coefficients

Prediction-inspired intelligent training for the development of c-RASAR models for organic skin sensitizers: Assessment of classification error rate from novel similarity coefficients

Contact Info

Product

Resources

About