Candidate Items Research Articles

P270 A global consensus on the definitions, diagnosis and management of fibrostenosing small bowel Crohn’s disease

Abstract Background Small bowel fibrostenotic strictures are common in patients with Crohn’s disease (CD). No global consensus recommendations on definitions, diagnosis and clinical management are available. Methods Several systematic reviews followed by a RAND/University of California Los Angeles appropriateness study on the definitions, diagnosis and clinical management of fibrostenosing CD in clinical practice were performed. A panel of 28 global experts and a patient representative were convened. They assessed a total of 152 candidate items. The items were subsequently evaluated for appropriateness. Results No accurate predictive biomarkers are available for naïve or anastomotic fibrostenosing strictures. Accurate diagnosis of fibrostenosing CD requires cross-sectional imaging which should evaluate bowel wall thickness, luminal narrowing and prestenotic dilatation. A potential inflammatory component should be assessed. Abdominal cross-sectional imaging was considered necessary prior to any treatment decision. The panel proposed an approach to medical, endoscopic, and surgical therapies (Figure 1 and Table 1). Technical characteristics for endoscopic balloon dilation and follow up strategies after successful dilation therapy were identified. Appropriateness, types and performance of different surgical approaches in various settings were evaluated. Conclusion This global consensus provides clinical guidance for the diagnostic and therapeutic management of patients with fibrostenotic CD.

P311 Reliability of computed tomography enterography for evaluation of stricturing Crohn’s disease: Development of a Crohn’s disease stricture index

Abstract Background Computed tomography enterography (CTE) is frequently used to assess stricturing Crohn’s disease (CD). However, the reliability of CTE items used for assessment has not been established and there is no validated radiologic stricture severity index. Methods We conducted a retrospective study of 43 patients with symptomatic, terminal ileal stricturing CD. Four radiologists assessed a comprehensive list of pre-defined radiologic items potentially associated with stricture severity in 48 CTE scans in two separate rounds in random order. Reliability was quantified using the intraclass correlation coefficient (ICC). Items with at least moderate (ICC≥0.41) inter-rater reliability that were correlated with a visual analogue scale of overall stricture severity were candidate items for the development of a CTE stricture index. Results Inter-rater reliability was almost perfect for assessment of luminal diameter of pre-stenotic dilation (ICC 0.817 [95% confidence interval 0.703, 0.878]), substantial for stricture length (ICC 0.628 [0.419, 0.797]), and moderate for stricture wall thickness (0.482 [0.330, 0.601]). Intra-rater reliability was almost perfect (ICC ≥ 0.850) for all three items. A stricture severity index was derived and was well calibrated (optimism-adjusted calibration slope = 1.009), with scores calculated as the sum of stricture length (cm) + luminal diameter of pre-stenotic dilation (mm) + 5 x stricture wall thickness (mm). Conclusion Stricture length, prestenotic dilation, and stricture wall thickness can be reliably assessed, and are component items for a novel CTE stricture index (CTE-SI) of stricture severity for CD-related terminal ileal strictures. Additional external index validation for clinical and research contexts is required.

Face and content validation of the amyotrophic lateral sclerosis-Bulbar dysfunction index (ALS-BDI).

Early detection and tracking of bulbar dysfunction in amyotrophic lateral sclerosis (ALS) are critical for directing management of the disease. Existing physiological assessments of bulbar dysfunction are often inaccessible and cost-prohibitive for clinical application. Existing clinical assessments are limited. The overall goal of our research is to develop a brief and reliable, clinician-administered assessment tool, the ALS Bulbar Dysfunction Index (ALS-BDI) to evaluate bulbar dysfunction. The aim of this study was to establish content and face validity of the ALS-BDI through item generation and reduction, including item scoring. The design of the ALS-BDI followed guidelines outlined by the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN). The design stage of the ALS-BDI involved two steps: (Step 1) the generation of candidate items from a literature review of commonly used clinical tools, and selection of items following a review of item reliability and item relevance and expert consensus; (Step 2) the assessment of their content and face validity via online survey feedback from experts (n = 35). The initial design was followed by a semi-structured cognitive interview with Speech-Language Pathologists (n = 5) to finalize a testable draft of the instrument. Two drafts of the ALS-BDI were developed. The first draft contained 48 items, after a review of existing clinical tools for their relevance to bulbar dysfunction in ALS. Of the 48 items, 35 items were retained after surveying experts and clinician users for their relevance, feasibility, interpretability, and appropriateness. The second draft of the ALS-BDI contained 37 items, due to one item splitting, based on users cognitive interviews. The ALS-BDI described in this study aims to provide a brief and reliable, clinician-administered assessment tool to evaluate bulbar dysfunction in patients with ALS. Future research will evaluate the psychometric properties of this tool including its reliability, validity, and responsiveness to change over time.

Disentangled Representation Learning for Recommendation.

There exist complex interactions among a large number of latent factors behind the decision making processes of different individuals, which drive the various user behavior patterns in recommender systems. These factors hidden in those diverse behaviors demonstrate highly entangled patterns, covering from high-level user intentions to low-level individual preferences. Uncovering the disentanglement of these latent factors can benefit in enhanced robustness, interpretability, and controllability during representation learning for recommendation. However, the large degree of entanglement within latent factors poses great challenges for learning representations that disentangle them, and remains largely unexplored in literature. In this paper, we present the SEMantic MACRo-mIcro Disentangled Variational Auto-Encoder (SEM-MacridVAE) model for learning disentangled representations from user behaviors, taking item semantic information into account. Our SEM-MacridVAE model achieves macro disentanglement by inferring the high-level concepts associated with user intentions (e.g., to buy a pair of shoes or a laptop) through a prototype routing mechanism, as well as capturing the individual preferences with respect to different concepts separately. The micro disentanglement is guaranteed through a micro-disentanglement regularizer stemming from an information-theoretic interpretation of VAEs, which forces each dimension of the representations to independently reflect an isolated low-level factor (e.g., the size or the color of a shirt). The semantic information including visual and categorical signals extracted from candidate items is utilized to further boost the recommendation performance of the proposed SEM-MacridVAE model. Empirical experiments demonstrate that our proposed approach is able to achieve significant improvement over the state-of-the-art baselines. We also show that the learned representations are interpretable and controllable, capable of potentially leading to a new paradigm for recommendation where users have fine-grained control over some target aspects of the recommendation candidates.

A Novel Scale to Assess Psychological Strategies in Explosive Ordnance Disposal Technicians.

This report describes the development and validation of the U.S. Navy Explosive Ordnance Disposal (EOD) Combat Mindset Scale-Training (CMS-T), a population-specific measure of psychological strategy use in EOD training environments. Scale items were developed by a working group composed of active-duty technicians from EOD Training and Evaluation Unit 1, Naval Health Research Center scientists, and a psychometrician. The working group developed 30 candidate items, which were administered to EOD accessions (new recruits), advanced students, and technicians (N = 164). Factor structure was explored with principal axis factoring and Varimax rotation with Kaiser normalization. Internal consistencies were established via Cronbach alpha, and convergent validity was evaluated with correlational and ANOVA models. Five internally stable subscales were derived from 19 essential items, explaining 65% of total variance. The subscales were named relaxation, attentional-emotional control (AEC), goal setting-visualization (GSV), internal dialogue (ID), and automaticity. The most frequently used strategies were GSV and ID. Expected relationships emerged between strategies, most notably AEC and mental health. The scale also differentiated between subgroups. The EOD CMS-T demonstrates a stable factor structure, internal reliability, and convergent validity. This study yields a valid, practical, and easily administered instrument to support EOD training and evaluation.

Recommendation model based on intention decomposition and heterogeneous information fusion.

In order to solve the problem of timeliness of user and item interaction intention and the noise caused by heterogeneous information fusion, a recommendation model based on intention decomposition and heterogeneous information fusion (IDHIF) is proposed. First, the intention of the recently interacting items and the users of the recently interacting candidate items is decomposed, and the short feature representation of users and items is mined through long-short term memory and attention mechanism. Then, based on the method of heterogeneous information fusion, the interactive features of users and items are mined on the user-item interaction graph, the social features of users are mined on the social graph, and the content features of the item are mined on the knowledge graph. Different feature vectors are projected into the same feature space through heterogeneous information fusion, and the long feature representation of users and items is obtained through splicing and multi-layer perceptron. The final representation of users and items is obtained by combining short feature representation and long feature representation. Compared with the baseline model, the AUC on the Last.FM and Movielens-1M datasets increased by 1.83 and 4.03 percentage points, respectively, the F1 increased by 1.28 and 1.58 percentage points, and the Recall@20 increased by 3.96 and 2.90 percentage points. The model proposed in this paper can better model the features of users and items, thus enriching the vector representation of users and items, and improving the recommendation efficiency.

Session-Based Recommendations for e-Commerce with Graph-Based Data Modeling

Conventional recommendation methods such as collaborative filtering cannot be applied when long-term user models are not available. In this paper, we propose two session-based recommendation methods for anonymous browsing in a generic e-commerce framework. We represent the data using a graph where items are connected to sessions and to each other based on the order of appearance or their co-occurrence. In the first approach, called Hierarchical Sequence Probability (HSP), recommendations are produced using the probabilities of items’ appearances on certain structures in the graph. Specifically, given a current item during a session, to create a list of recommended next items, we first compute the probabilities of all possible sequential triplets ending in each candidate’s next item, then of all candidate item pairs, and finally of the proposed item. In our second method, called Recurrent Item Co-occurrence (RIC), we generate the recommendation list based on a weighted score produced by a linear recurrent mechanism using the co-occurrence probabilities between the current item and all items. We compared our approaches with three state-of-the-art Graph Neural Network (GNN) models using four session-based datasets one of which contains data collected by us from a leather apparel e-shop. In terms of recommendation effectiveness, our methods compete favorably on a number of datasets while the time to generate the graph and produce the recommendations is significantly lower.

Typical Hits, Grams, or Joints: Evaluating Cannabis Survey Measurement Strategies for Quantifying Consumption.

Aim: Standardized survey measures that capture diverse cannabis consumption patterns are needed to inform public health and policy. Our team is developing a flexible, personalized, low-burden survey item inventory to measure cannabis use patterns and estimate milligrams of THC (mgTHC) consumption in large samples. This study aimed to identify measurement gaps and analysis implications associated with an initial pool of candidate items that assessed use of cannabis flower and concentrate products (smoked and/or vaporized). Methods: Adult cannabis consumers (n=4247) completed an online survey assessing cannabis use frequency, quantity, product types, product potencies (%THC), and methods of administration. Participants chose to report their consumption quantities using one of three units: "hits per day," "grams per week," or "joints per week." Respondents also indicated whether their past 7-day consumption pattern represented their typical pattern. Results: Eighty-one percent had used cannabis daily in the past week. Thirty-two percent, 53%, and 15% chose to report flower and concentrate consumption quantity in hits, grams, and joints, respectively. Approximately 80-90% of responses for the number of hits, grams, and joints consumed were less than the maximum response option-suggesting that response options captured the full range of potential cannabis consumption behaviors. Those who chose grams or joints units were generally more likely to endorse higher risk cannabis use (e.g., morning use, high %THC products) in the past week than those who chose the hits unit (adjusted Odds Ratio range: 1.2-3.9). Among those who reported that the past week represented their typical behavior (83%), past 30-day and past 7-day frequencies were highly correlated (Spearman's Rho=0.77)-supporting the feasibility of using lower burden "typical week" items to extrapolate patterns beyond a 1-week time frame. Conclusion: Results from this online convenience sample of frequent cannabis consumers suggest that the current items yield coherent and expected response patterns. Although additional testing is required, a standardized, flexible survey instrument for large-scale assessment of cannabis patterns and calculation of mgTHC seems within reach.

Key items for reports of primary care research: an international Delphi study

ObjectiveReporting guidelines can improve dissemination and application of findings and help avoid research waste. Recent studies reveal opportunities to improve primary care (PC) reporting. Despite increasing numbers of guidelines, none...

Development of a core outcome set for quality of life for adults with drug-resistant epilepsy: A multistakeholder Delphi consensus study.

In 2017, the American Academy of Neurology (AAN) convened the AAN Quality Measurement Set working group to define the improvement and maintenance of quality of life (QOL) as a key outcome measure in epilepsy clinical practice. A core outcome set (COS), defined as an accepted, standardized set of outcomes that should be minimally measured and reported in an area of health care research and practice, has not previously been defined for QOL in adult epilepsy. A cross-sectional Delphi consensus study was employed to attain consensus from patients and caregivers on the QOL outcomes that should be minimally measured and reported in epilepsy clinical practice. Candidate items were compiled from QOL scales recommended by the AAN 2017 Quality Measurement Set. Inclusion criteria to participate in the Delphi study were adults with drug-resistant epilepsy diagnosed by a physician, no prior diagnosis of psychogenic nonepileptic seizures or a cognitive and/or developmental disability, or caregivers of patients meeting these criteria. A total of 109 people satisfied inclusion/exclusion criteria and took part in Delphi Round 1 (patients, n= 95, 87.2%; caregivers, n= 14, 12.8%), and 55 people from Round 1 completed Round 2 (patients, n= 43, 78.2%; caregivers, n= 12, 21.8%). One hundred three people took part in the final consensus round. Consensus was attained by patients/caregivers on a set of 36 outcomes that should minimally be included in the QOL COS. Of these, 32 of the 36 outcomes (88.8%) pertained to areas outside of seizure frequency and severity. Using patient-centered Delphi methodology, this study defines the first COS for QOL measurement in clinical practice for adults with drug-resistant epilepsy. This set highlights the diversity of factors beyond seizure frequency and severity that impact QOL in epilepsy.

2022 American College of Rheumatology/EULAR Classification Criteria for Giant Cell Arteritis.

To develop and validate updated classification criteria for giant cell arteritis (GCA). Patients with vasculitis or comparator diseases were recruited into an international cohort. The study proceeded in 6 phases: 1) identification of candidate items, 2) prospective collection of candidate items present at the time of diagnosis, 3) expert panel review of cases, 4) data-driven reduction of candidate items, 5) derivation of a points-based risk classification score in a development data set, and 6) validation in an independent data set. The development data set consisted of 518 cases of GCA and 536 comparators. The validation data set consisted of 238 cases of GCA and 213 comparators. Age ≥50 years at diagnosis was an absolute requirement for classification. The final criteria items and weights were as follows: positive temporal artery biopsy or temporal artery halo sign on ultrasound (+5); erythrocyte sedimentation rate ≥50 mm/hour or C-reactive protein ≥10 mg/liter (+3); sudden visual loss (+3); morning stiffness in shoulders or neck, jaw or tongue claudication, new temporal headache, scalp tenderness, temporal artery abnormality on vascular examination, bilateral axillary involvement on imaging, and fluorodeoxyglucose-positron emission tomography activity throughout the aorta (+2 each). A patient could be classified as having GCA with a cumulative score of ≥6 points. When these criteria were tested in the validation data set, the model area under the curve was 0.91 (95% confidence interval [95% CI] 0.88-0.94) with a sensitivity of 87.0% (95% CI 82.0-91.0%) and specificity of 94.8% (95% CI 91.0-97.4%). The 2022 American College of Rheumatology/EULAR GCA classification criteria are now validated for use in clinical research.

2022 American College of Rheumatology/EULAR Classification Criteria for Takayasu Arteritis.

To develop and validate new classification criteria for Takayasu arteritis (TAK). Patients with vasculitis or comparator diseases were recruited into an international cohort. The study proceeded in 6 phases: 1) identification of candidate criteria items, 2) collection of candidate items present at diagnosis, 3) expert panel review of cases, 4) data-driven reduction of candidate items, 5) derivation of a points-based classification score in a development data set, and 6) validation in an independent data set. The development data set consisted of 316 cases of TAK and 323 comparators. The validation data set consisted of an additional 146 cases of TAK and 127 comparators. Age ≤60 years at diagnosis and imaging evidence of large-vessel vasculitis were absolute requirements to classify a patient as having TAK. The final criteria items and weights were as follows: female sex (+1), angina (+2), limb claudication (+2), arterial bruit (+2), reduced upper extremity pulse (+2), reduced pulse or tenderness of a carotid artery (+2), blood pressure difference between arms of ≥20 mm Hg (+1), number of affected arterial territories (+1 to +3), paired artery involvement (+1), and abdominal aorta plus renal or mesenteric involvement (+3). A patient could be classified as having TAK with a cumulative score of ≥5 points. When these criteria were tested in the validation data set, the model area under the curve was 0.97 (95% confidence interval [95% CI] 0.94-0.99) with a sensitivity of 93.8% (95% CI 88.6-97.1%) and specificity of 99.2% (95% CI 96.7-100.0%). The 2022 American College of Rheumatology/EULAR classification criteria for TAK are now validated for use in research.

Development of an inflammatory bowel disease (IBD) Patient-Reported Experience Measure (PREM): A patient-led consensus work and 'think aloud' study for a quality improvement programme.

Patient-Reported Experience Measures (PREMs) are key in improving healthcare quality, but no PREM exists for inflammatory bowel disease (IBD). This study aimed to co-produce a PREM with IBD service users for IBD service evaluation and quality improvement programme. A pool of 75 items was drawn from published survey instruments covering interactions with services and aspects of living with IBD. In Stage 1, during two workshops, eight expert service users reduced candidate items through a ranked-choice voting exercise and suggested further items. During Stage 2, 18 previously uninvolved people with IBD assessed the face and content validity of the candidate items in 'Think Aloud' interviews. During two final workshops (Stage 3), the expert service users removed, modified and added items based on the interview findings to produce a final version of the PREM. Stage 1 generated a draft working PREM mapped to the following four domains: Patient-Centred Care; Quality; Accessibility; Communication and Involvement. The PREM included a set of nine items created by the expert group which shifted the emphasis from 'self-management' to 'living with IBD'. Stage 2 interviews showed that comprehension of the PREM was very good, although there were concerns about the wording, IBD-relevance and ambiguity of some items. During the final two workshops in Stage 3, the expert service users removed 7 items, modified 15 items and added seven new ones based on the interview findings, resulting in a 38-item PREM. This study demonstrates how extensive service user involvement can inform PREM development. Patients were involved as active members of the research team and as research participants to co-produce and validate a PREMfor IBD services. In Stage 1, eight expert service users ('the expert group') reduced candidate items for the PREM through a voting exercise and suggested new items. During Stage 2, 18 previously uninvolved people with IBD (the 'think aloud' participants) assessed the validity of the candidate items in 'Think Aloud' interviews as research participants. In Stage 3, the expert group removed, changed and added items based on the interview findings to produce a final version of the 38-item PREM. This study shows how service user involvement can meaningfully inform PREM development.

Qualitative development of the PROMIS Profile v1.0-Familial Chylomicronemia Syndrome (FCS) 28

PurposeFamilial chylomicronemia syndrome (FCS) is a rare genetic disorder characterized by high triglyceride levels, significant disease burden, and negative impacts on health-related quality of life. This project aimed to create a PROMIS-based patient-reported outcome measure that represents valid and important concerns for patients with FCS.MethodsWe reviewed the literature and data from a previous qualitative study of FCS to identify key FCS symptoms and impacts, which were mapped to PROMIS domains to create a pool of eligible items. Candidate items were reduced per expert feedback and patients with FCS completed cognitive interviews to confirm content validity and measure content.ResultsLiterature and qualitative data review identified ten key symptoms and 12 key impacts of FCS, including abdominal pain, fatigue, difficulty thinking, and worry about pancreatitis attacks. We identified 96 items primarily from PROMIS, supplemented with items from the Quality of Life in Neurological Disorders™ (Neuro-QoL™) and the Functional Assessment of Chronic Illness Therapy (FACIT) measurement systems. This pool was reduced to 32 candidate items, which were assessed via cognitive interviews with eight participants with FCS. Cognitive interview results and additional expert feedback led to the removal of four items and finalization of the PROMIS Profile v1.0—familial chylomicronemia syndrome (FCS) 28.ConclusionsThe PROMIS Profile v1.0—familial chylomicronemia syndrome (FCS) 28 provides strong content validity for assessing quality of life among patients with FCS. The benefits of PROMIS, including norm-referenced mean values for each measure, will facilitate comparison of patients with FCS to other clinical populations.

Development of an EORTC Item Bank for Computer-Adaptive Testing of Nausea and Vomiting

ObjectivesNausea and vomiting (NV) remain common cancer symptoms and frequent side effects of anticancer therapies despite available antiemetics. They can lead to treatment disruption and discontinuation. NV is an important patient reported outcome in oncology. This study aimed to build an item bank for computer-adaptive testing (CAT) based on NV questions in the European Organisation for Research for Treatment of Cancer, Quality of Life for Cancer Patients (EORTC QLQ-C30) questionnaire and complete the first three phases of development as described in the EORTC Quality of Life Group guidelines. Data SourcesThe development followed a standard procedure. The three phases include conceptualization and literature search (phase 1); item classification, selection, formulation and rating, and expert evaluations (phase 2); and patient pretesting (phase 3). The literature search resulted in a preliminary list of 115 items. Following classification, formulation, and rating, 21 candidate items adhered to the QLQ-C30 format. Evaluation by experts (n = 11) from five countries and patients (n = 31) pretesting in Denmark, Poland, and the UK lead to a final list of 20 items. ConclusionThe selection, development, and refining of NV items have been described. The nature of this testing ensures an initial CAT item bank that after field testing (phase 4) and psychometric analysis is expected to provide a precise and efficient NV measurement while still being comparable to the original QLQ-C30 scale. Implications for Nursing PracticeAccess to reliable tools that facilitate NV comprehensive assessment is an important issue for nurses caring for patients with cancer. This CAT item bank is meant to support clinical decisions when all phases of testing are completed.

A Web-Based Prostate Cancer-Specific Holistic Needs Assessment (CHAT-P): Multimethod Study From Concept to Clinical Practice.

Men with prostate cancer experience immediate and long-term consequences of the disease and its treatment. They require both long-term monitoring for recurrence or progression and follow-up to identify and help manage psychosocial and physical impacts. Holistic Needs Assessment aims to ensure patient-centered continuing cancer care. However, paper-based generic tools have had limited uptake within cancer services, and there is little evidence of their impact. With the expansion of remote methods of care delivery and to enhance the value of generic tools, we developed a web-based Composite Holistic Needs Assessment Adaptive Tool-Prostate (CHAT-P) specifically for prostate cancer. This paper described the context, conceptual underpinning, and approach to design that informed the development of CHAT-P, starting from the initial concept to readiness for deployment. Through this narrative, we sought to contribute to the expanding body of knowledge regarding the coproduction process of innovative digital systems with potential for enhanced cancer care delivery. The development of CHAT-P was guided by the principles of coproduction. Men with prostate cancer and health care professionals contributed to each stage of the process. Testing was conducted iteratively over a 5-year period. An initial rapid review of patient-reported outcome measures identified candidate items for inclusion. These items were categorized and allocated to overarching domains. After the first round of user testing, further items were added, improvements were made to the adaptive branching system, and response categories were refined. A functioning version of CHAT-P was tested with 16 patients recruited from 3 outpatient clinics, with interviewers adopting the think-aloud technique. Interview transcripts were analyzed using a framework approach. Interviews and informal discussions with health care professionals informed the development of a linked care plan and clinician-facing platform, which were incorporated into a separate feasibility study of digitally enhanced integrated cancer care. The findings from the interview study demonstrated the usability, acceptability, and potential value of CHAT-P. Men recognized the benefits of a personalized approach and the importance of a holistic understanding of their needs. Preparation for the consultation by the completion of CHAT-P was also recognized as empowering. The possible limitations identified were related to the importance of care teams responding to the issues selected in the assessment. The subsequent feasibility study highlighted the need for attention to men's psychological concerns and demonstrated the ability of CHAT-P to capture red flag symptoms requiring urgent investigation. CHAT-P offers an innovative means by which men can communicate their concerns to their health care teams before a physical or remote consultation. There is now a need for a full evaluation of the implementation process and outcomes where CHAT-P is introduced into the clinical pathway. There is also scope for adapting the CHAT-P model to other cancers.

The Muscularity Bias Internalization Scale: Development and initial validation in Chinese adult men

The present study developed the Muscularity Bias Internalization Scale (MBIS), a measure of muscularity bias internalization, which assesses the extent to which an individual endorses muscularity-based stereotypes and engages in negative self-evaluations due to muscularity. In Study 1, a pool of candidate items was developed. By using exploratory factor analysis (EFA) on a sample of 300 Chinese adult men (Mage = 29.98 years, SD = 7.81), 14 items from the item pool were retained. The 14-item MBIS contained three factors. In Study 2, a second sample of 300 Chinese adult men (Mage = 29.50 years, SD = 7.50) was used to confirm the factor structure of the MBIS, as well as to examine its reliability and validity. Confirmatory factor analysis (CFA) further confirmed the three-factor structure of the MBIS with a good model fit. The MBIS showed high internal consistency reliability and high two-week test-retest reliability. The MBIS also showed good construct validity with measures of drive for muscularity, muscularity-oriented disordered eating, and muscle dysmorphia symptoms. The newly developed MBIS may help understand the role of muscularity bias internalization in the development and maintenance of muscularity-oriented body image and muscularity-oriented disordered eating.

Improve the Assessment Model of Personnel Develop Level in Higher Education Based on Machine Learning

In order to improve the ability to manage the quality constraint index parameters of higher education talent training, an evaluation model of higher education talent training quality based on improved machine learning is proposed. Combined with the characteristic analysis method of higher education talent training quality constraint index parameters, an autocorrelation feature matching model of higher education talent training quality constraint index parameters is established, and fuzzy association rule scheduling is used to realize the feature extraction and fuzzy clustering of higher education talent training quality constraint index parameters. The mutual information coupling parameter analysis of the quality constraint index parameters of higher education talent training, combined with the generalized association rules and periodic association rules, realizes the feature clustering and global candidate item set analysis of the quality constraint index parameters of higher education talent training and realizes the dynamic evaluation of the quality of higher education talent training according to the machine learning analysis results. The test results show that this method has good clustering of data characteristics, good reliability of the evaluation process, and good convergence of machine learning process, which improves the dynamic and quantitative management ability of talent training quality in higher education.

REDRL: A review-enhanced Deep Reinforcement Learning model for interactive recommendation

Recent advances in interactive recommender systems (IRS) have received wide attention due to its flexible recommendation strategy and optimization for users’ long-term utility. Considering this interaction paradigm of IRS, researchers have made some attempts to incorporate reinforcement learning (RL) models into IRS, because of the excellent ability of RL in long-term optimizing and decision making. However, data sparsity is an intractable problem most IRS urgently need to address. Although a small amount of work has exploited reviews to address data sparsity, they ignored the varying importance of items for modeling the user. In addition, most existing RL-based approaches suffer from decision-making difficulties when the action space becomes large. To solve above problems, in this work, we present a Review-enhanced Deep Reinforcement Learning model (REDRL) for interactive recommendation. Specifically, we utilize text reviews, combined with a pretrained review representation model to acquire item review-enhanced embedding representations. Then we formalize the recommendation problem as a Markov Decision Process (MDP), and exploit deep reinforcement learning (DRL) to model the interactive recommendation. Notably, we introduce a multi-head self-attention technique to capture distinct importance of various items in the sequence behavior, which is overlooked by existing work when modeling the user preference. In this way, we can model long-term dynamic preferences of users accurately and discriminately for comprehensive interactive recommendation. Moreover, we subtly combine the semantic structure information in the user–item bipartite graph with meta-paths in heterogeneous information networks (HIN), to filter some irrelevant items and obtain candidate items dynamically. By this means, the size of the discrete action space is effectively reduced from a new anger. The experimental results based on three benchmark datasets demonstrate the efficiency of our method with significant improvement over state-of-the-art.

Performance of re-ranking techniques used for recommendation method to the user CF- Model

The recent research work for addressed to the aims at a spectrum of item ranking techniques that would generate recommendations with far more aggregate variability across all users while retaining comparable levels of recommendation accuracy. Individual users and companies are increasingly relying on recommender systems to provide information on individual suggestions. The recommended technologies are becoming increasingly efficient because they are focusing on scalable sorting-based heuristics that make decisions based solely on "local" data (i.e., only on the candidate items of each user) rather than having to keep track of "national" data, such as items have been all user recommended at the time. The real-world rating datasets and various assessments to be the prediction techniques and comprehensive empirical research consistently demonstrate the proposed techniques' diversity gains. Although the suggested approaches have primarily concentrated on improving recommendation accuracy, other critical aspects of recommendation quality, such as recommendation delivery, have often been ignored.