Latent Preferences Research Articles

Inverting Cognitive Models With Neural Networks to Infer Preferences From Fixations.

Inferring an individual's preferences from their observable behavior is a key step in the development of assistive decision-making technology. Although machine learning models such as neural networks could in principle be deployed toward this inference, a large amount of data is required to train such models. Here, we present an approach in which a cognitive model generates simulated data to augment limited human data. Using these data, we train a neural network to invert the model, making it possible to infer preferences from behavior. We show how this approach can be used to infer the value that people assign to food items from their eye movements when choosing between those items. We demonstrate first that neural networks can infer the latent preferences used by the model to generate simulated fixations, and second that simulated data can be beneficial in pretraining a network for predicting human-reported preferences from real fixations. Compared to inferring preferences from choice alone, this approach confers a slight improvement in predicting preferences and also allows prediction to take place prior to the choice being made. Overall, our results suggest that using a combination of neural networks and model-simulated training data is a promising approach for developing technology that infers humanpreferences.

Investigating aquaculture farmers' preferences for multiobjective green production schemes: A discrete choice experiment in China

In recent years, the objectives of green agricultural development have increasingly diversified, incorporating food safety, pollution control, and resource conservation. This diversification has further heightened the potential for conflicts between policy objectives and the individual interests of farmers. Understanding farmer trade-offs with respect to multiobjective green production (MOGP) policies is crucial for the comprehensive implementation of agricultural green development policies. The green and healthy aquaculture action (GHAA) encompassing multiple green technology objectives promoted by the Chinese government provides a favorable policy context. On the basis of the policy background of the GHAA, this study utilized the choice experiment method (CEM) to explore the preferences of aquaculture farmers for various green technology objectives and supporting measures. The results indicate that aquaculture farmers are willing to make trade-offs among various green technology objectives with the support of specific measures to transform traditional production methods. Farmers have varying preferences for different green technology objectives. Overall, they tend to prefer implementing compound feed utilization but are averse to implementing aquaculture effluent treatment. In terms of support measures, providing technical guidance is more effective than providing aquaculture insurance in incentivizing farmers to participate in MOGP schemes. Further analysis indicates that there are three latent preference classes, which differ in their preferences for green technology objectives and support measures. Age, gender, aquaculture area, risk preferences, and participation in cooperatives explain the probability of preference class membership.

HyperCLR: A Personalized Sequential Recommendation Algorithm Based on Hypergraph and Contrastive Learning

Sequential recommendations aim to predict users’ next interactions by modeling their interaction sequences. Most existing work concentrates on user preferences within these sequences, overlooking the complex item relationships across sequences. Additionally, these studies often fail to address the diversity of user interests, thus not capturing their varied latent preferences effectively. To tackle these problems, this paper develops a novel recommendation algorithm based on hypergraphs and contrastive learning named HyperCLR. It dynamically incorporates the time and location embeddings of items to model high-order relationships in user preferences. Moreover, we developed a graph construction approach named IFDG, which utilizes global item visit frequencies and spatial distances to discern item relevancy. By sampling subgraphs from IFDG, HyperCLR can align the representations of identical interaction sequences closely while distinguishing them from the broader global context on IFDG. This approach enhances the accuracy of sequential recommendations. Furthermore, a gating mechanism is designed to tailor the global context information to individual user preferences. Extensive experiments on Taobao, Books and Games datasets have shown that HyperCLR consistently surpasses baselines, demonstrating the effectiveness of the method. In particular, in comparison to the best baseline methods, HyperCLR demonstrated a 29.1% improvement in performance on the Taobao dataset.

A tensor recommendation method based on HMM network and meta-path

The main approach to capturing user latent preferences in recommendation systems (RS) is through high-order tensor decomposition and the deep-walk method. Several key issues, if solved, could improve the performance of RS. These include enforcing the interpretation of RS in the context of sparse data completion, cold start, and interpretability, mining user latent preferences with a tensor constructed from a user-item rating matrix (RM) and a preference match mechanism based on K-nearest neighbor (KNN) similar users. In this paper, a method that integrates a hidden Markov model, meta-path, and third-order tensor (HMM-MP-TOT) is proposed. An HMM, based on the user-item RM and latent preferences from KNN users is constructed. Subsequently, the Viterbi and deep-walk methods are used to obtain a series of user-item two-dimensional MPs. Then, truncated − singular value decomposition (t-SVD) is applied to a user-item-KNN third-order tensor to obtain a better recommendation result. On average, HMM-MP-TOT obtains 94.7% precision, 80.2% recall, and 96.4% diversity.

Measuring socially appropriate social preferences

This paper extends the literature on structural estimation of social preferences to account for the desire to adhere to social norms and hide one's true intentions via moral wiggle room. We conduct an experiment to test whether accounting for normatively appropriate behavior allows us to distinguish between preference types who care about outcomes versus adhering to social norms and whether the introduction of moral wiggle room undermines the stability of social preference estimates. We find that social preference estimates are remarkably robust to the inclusion of moral wiggle room. However, the representative agent is strongly motivated by norms and failing to account for this motive in our model causes us to overestimate how much agents care about helping those who are worse off. Using finite mixture models to endogenously identify latent preference types, we replicate previous work finding that the majority of subjects can be classified as strong or moderate altruists when normative concerns are not considered. Accounting for the normative appropriateness of decisions when categorizing participants, however, reveals different motives across types: strong altruists are only marginally concerned with norms while the moderate altruists are highly sensitive to them and, once norms are taken into account, don't care at all about the outcomes of others. Our results thus recast the prior findings in a new light. Rather than the two most common types being strong altruists and moderate altruists, we find that they are better described as strong altruists and norm followers.

MOOCs video recommendation using low-rank and sparse matrix factorization with inter-entity relations and intra-entity affinity information

PurposeThe serious information overload problem of MOOCs videos decreases the learning efficiency of the students and the utilization rate of the videos. There are two problems worthy of attention for the matrix factorization (MF)-based video learning resource recommender systems. Those methods suffer from the sparsity problem of the user-item rating matrix, while side information about user and item is seldom used to guide the learning procedure of the MF. MethodTo address those two problems, we proposed a new MOOCs video resource recommender LSMFERLI based on Low-rank and Sparse Matrix Factorization (LSMF) with the guidance of the inter-Entity Relations and intra-entity Latent Information of the students and videos. Firstly, we construct the inter-entity relation matrices and intra-entity latent preference matrix for the students. Secondly, we construct the inter-entity relation matrices and intra-entity affinity matrix for the videos. Lastly, with the guidance of the inter-entity relation and intra-entity affinity matrices of the students and videos, the student-video rating matrix is factorized into a low-rank matrix and a sparse matrix by the alternative iteration optimization scheme. ConclusionsExperimental results on dataset MOOCcube indicate that LSMFERLI outperforms 7 state-of-the-art methods in terms of the HR@K and NDCG@K(K = 5,10,15) indicators increased by an average of 20.6 % and 21.0 %, respectively.

How to overcome rent seeking in Tanzania’s skills sector? Exploring feasible reforms through discrete choice experiments

Skills gaps and mismatches are widely documented as a hindrance to inclusive structural transformation across developing countries, especially in Africa. What is often overlooked, however, is the fact that skills development is a complex political economy process challenged by institutional and financing problems on the supply side, and inadequate demand, that is, a shortage of firms that can organise skilled labour and provide on-the-job training effectively. In such adverse contexts, rent seeking and corruption may arise from conflicting objectives, trade-offs and mis-aligned incentives among stakeholders – public sector skills providers and firms. With a focus on Tanzania, we (i) analyse the incentive structures underlying such rule-breaking behaviours and processes, and (ii) empirically test alternative institutional design strategies that would better align the interests of different stakeholders towards improved skills development outcomes. Building on over 30 in-depth stakeholder interviews in 2018, we conducted three Discrete Choice Experiments with over 200 firms to test the feasibility of different incentive packages in 2019. Our main hypothesis is that the successful re-alignment of stakeholders’ incentives must consider both the different and potentially conflicting objectives of public training institutions and the heterogeneity in skills needs and capabilities of different types of firms. We uncover latent preference structures differentiated by observable firm characteristics, most strongly by differences in technical capabilities, existing training provision and firm size. We conclude advancing an evidence-based tailored skills policy reform.

Environmental benefits and supply dynamics of electric vehicles sharing: From a systematic perspective of transportation structure and trip purposes

This study elucidates the environmental benefits and the supply dynamics of electric vehicle sharing (EVS) across various trip modes and purposes. A comprehensive survey involving 1,784 participants is carried out to gauge travelers’ latent preferences to EVS. An integrated model of the structural equation modeling (SEM) and the partial proportional odds model (PPOM) is developed to identify the interactive relationships between EVS and other trip modes. A before-and-after analysis is conducted to unveil the environmental benefits of EVS across different trip modes and purposes, based on the Theory of Planned Behavior (TPB) and the Technology Acceptance Model (TAM). In particular, marginal effect analysis is performed to assess the environmental benefits and supply dynamics as EVS adoption levels vary. The findings highlight a remarkable reduction in energy consumption and CO2 emissions in Beijing due to EVS operations, amounting to reductions of 168,000 kgce and 406,000 kgCO2, respectively. This study provides policymakers with insights into EVS’s environmental potentials and optimization recommendations.

Modeling Urban Railway Passengers’ Willingness to Wait Based on Latent Class Analysis

Overcrowding in confined spaces, like inside trains, is a social issue that, in the long-term, could lead to accidents. To alleviate congestion in trains, a reduction in train headways is planned, made possible by technological advances such as the introduction of advanced autonomous trains (e.g., 60-s headway). However, simply reducing a train’s headway may be challenging because this does not consider the travel behavior of passengers who engage in the disruptive act of excessively boarding trains, thereby causing delays in their departure. Furthermore, the behavior of passengers may vary depending on an individual’s latent preferences, however, there is insufficient research that reflects these latent preferences. In this study, a survey was conducted with 971 urban rail passengers in the Seoul Metropolitan Area of Korea to identify their latent preferences using latent class analysis, resulting in a four-type classification: a congestion avoidance type, a time-sensitive type, a subway-preference type, and a type that does not prefer public transportation. Based on this, willingness to wait (WTW; i.e., waiting for the next less crowded train to reduce passenger discomfort) according to their latent preferences, was calculated. WTW was found to increase as the headway of the train was reduced. Choice modeling was conducted based on variables, including socioeconomic variables, to suggest effects related to WTW. The results indicated demand dispersion effects according to the headway reduction at peak times for various types of people. This study’s findings could be utilized to offer less crowded and safer train operations.

Uncovering the Latent Preferences of Slovenia’s Private Forest Owners in the Context of Enhancing Forest Ecosystem Services through a Hypothetical Scheme

Background and objectives: Successful policy implementation relies on understanding stakeholders’ willingness to contribute to policy goals. The EU Green Deal, with strategies on forests, biodiversity, and the bioeconomy, also depends on the performance of the forestry sector, including a significant portion of privately owned forests. Materials and methods: We conducted a nationwide survey among a sample of 341 private forest owners in Slovenia (total population of 424,086). The online questionnaire had three sections: (1) knowledge and priorities regarding ecosystem services and the bioeconomy, (2) a discrete choice experiment for eliciting preferences concerning the implementation of activities supporting strategic goals originating from EU Green Deal policies, and (3) socio-economic data and future forest management objectives. Results: The results indicated heterogeneity in preferences for performing activities on private forest lands to enhance specific forest ES that contribute to policy goals. More than half of the respondents (57.6%) exhibited a reluctance to implement activities and were skeptical of higher compensation payments, whereas the rest expressed an inclination towards changing their forest management. Conclusions: Slovenia’s private forest owners appear to be heterogeneous in their willingness to participate in a hypothetical ES enhancement scheme that could contribute to some EU Green Deal goals. Policymakers must recognize intrinsic motives and social norms that affect the willingness of forest owners to be engaged to increase the acceptance of solutions.

Patients accept therapy using embryonic stem cells for Parkinson’s disease: a discrete choice experiment

BackgroundNew disease-modifying ways to treat Parkinson’s disease (PD) may soon become a reality with intracerebral transplantation of cell products produced from human embryonic stem cells (hESCs). The aim of this study was to assess what factors influence preferences of patients with PD regarding stem-cell based therapies to treat PD in the future.MethodsPatients with PD were invited to complete a web-based discrete choice experiment to assess the importance of the following attributes: (i) type of treatment, (ii) aim of treatment, (iii) available knowledge of the different types of treatments, (iv) effect on symptoms, and (v) risk for severe side effects. Latent class conditional logistic regression models were used to determine preference estimates and heterogeneity in respondents’ preferences.ResultsA substantial difference in respondents’ preferences was observed in three latent preference patterns (classes). “Effect on symptoms” was the most important attribute in class 1, closely followed by “type of treatment,” with medications as preferred to other treatment alternatives. Effect on symptoms was also the most important attribute in class 2, with treatment with hESCs preferred over other treatment alternatives. Likewise for class 3, that mainly focused on “type of treatment” in the decision-making. Respondents’ class membership was influenced by their experience in treatment, side effects, and advanced treatment therapy as well as religious beliefs.ConclusionsMost of the respondents would accept a treatment with products emanating from hESCs, regardless of views on the moral status of embryos. Preferences of patients with PD may provide guidance in clinical decision-making regarding treatments deriving from stem cells.

Not going out: television’s impacts on Britain’s commercial entertainment industries and popular leisure during the 1950s

ABSTRACT The 1950s was a pivotal decade for Britain’s entertainment industries, with the rapid diffusion of television and sharp declines for hitherto dominant urban venue entertainments. This had important social consequences, including the acceleration of the trend from community-based socialising to more sedentary, family-based, entertainment – the last essential component of the ‘industrialisation of the home’. However, the disruptive impact of television varied considerably among different incumbent urban entertainments, with variety theatre and cinema facing catastrophic declines, while spectator sports and dance halls continued to flourish. This article examines television’s differential impact on incumbent entertainments using a variety of new sources, including Customs and Excise data; unpublished government social surveys; and trade sources. The differential impact of television on incumbent entertainments can be largely explained by the degree of ‘commitment’ demanded of consumers for different leisure activities; the degree to which television was a strong substitute; the presence of addictive elements (gambling); and the extent to which the activity appealed to a youth audience. However, the rapid collapse of variety theatre and cinema can only be fully explained by television enabling strong latent preferences for commercial entertainment in the home, which were now satisfied by television.

When firms may benefit from sticking with an old technology

AbstractResearch SummaryHow should firms respond to technological discontinuities in order to achieve greater performance? In contrast to most studies that advocate a timely transition from the old to the new technology, this paper posits that in markets where a discontinuous technology exposes customers' latent preference heterogeneity for certain old technology attributes, firms may ultimately experience a performance surge by adhering to the old technology during technological change. Explicitly, I theorize a U‐shaped relationship within such a market between competitors' increasing adoption of the new technology and the performance of firms that stick with the old technology. This prediction is thoroughly examined using comprehensive data from the traditional Chinese medicine industry in China during the 1990s and receives robust empirical support.Managerial SummaryIn some markets, the rise of a discontinuous technology, besides posing a substitute threat to the old technology, further exposes niche segments where customers continue to favor the old technology. This paper predicts that within such a market, as competitors increasingly adopt the new technology for varied motives, firms sticking with the old technology may see their performance declining before rebounding and potentially reaching new heights. Analyses using archival data from the traditional Chinese medicine industry in China during the 1990s provide robust support for this prediction. The arguments and findings of this paper offer an “existence proof” that when confronted with a technological discontinuity, adhering to the old technology may also represent an effective strategy that ultimately improves firm performance.

A Hybrid Recommender System Based on Autoencoder and Latent Feature Analysis.

A recommender system (RS) is highly efficient in extracting valuable information from a deluge of big data. The key issue of implementing an RS lies in uncovering users' latent preferences on different items. Latent Feature Analysis (LFA) and deep neural networks (DNNs) are two of the most popular and successful approaches to addressing this issue. However, both the LFA-based and the DNNs-based models have their own distinct advantages and disadvantages. Consequently, relying solely on either the LFA or DNN-based models cannot ensure optimal recommendation performance across diverse real-world application scenarios. To address this issue, this paper proposes a novel hybrid recommendation model that combines Autoencoder and LFA techniques, termed AutoLFA. The main idea of AutoLFA is two-fold: (1) It leverages an Autoencoder and an LFA model separately to construct two distinct recommendation models, each residing in a unique metric representation space with its own set of strengths; and (2) it integrates the Autoencoder and LFA model using a customized self-adaptive weighting strategy, thereby capitalizing on the merits of both approaches. To evaluate the proposed AutoLFA model, extensive experiments on five real recommendation datasets are conducted. The results demonstrate that AutoLFA achieves significantly better recommendation performance than the seven related state-of-the-art models.

Disentangled CVAEs with Contrastive Learning for Explainable Recommendation

Modern recommender systems are increasingly expected to provide informative explanations that enable users to understand the reason for particular recommendations. However, previous methods struggle to interpret the input IDs of user--item pairs in real-world datasets, failing to extract adequate characteristics for controllable generation. To address this issue, we propose disentangled conditional variational autoencoders (CVAEs) for explainable recommendation, which leverage disentangled latent preference factors and guide the explanation generation with the refined condition of CVAEs via a self-regularization contrastive learning loss. Extensive experiments demonstrate that our method generates high-quality explanations and achieves new state-of-the-art results in diverse domains.

Is it more important to increase carbon sequestration, biodiversity, or jobs? A case study of citizens' preferences for forest policy in Finland

Forest management plays an important role in mitigating the most serious threats to the well-being of current and future generations: climate change and biodiversity loss. We examine citizens' preferences for alternative forest management policy goals, including the provision of carbon sinks, biodiversity, and forest sector jobs, using a discrete choice experiment in Finland. We explore preference heterogeneity using latent class multinomial logit models finding the highest willingness to pay for biodiversity-oriented forest policy. Comparing three forest policy scenarios, focusing on “Biodiversity,” “Climate,” or “Bioeconomy,” citizens preferred the overall environmental goals over economic ones. We observe considerable heterogeneity in respondents' preferences identifying five latent preference classes. The results provide valuable background information on citizens' values and preferences on tradeoffs built into future revisions of forest policy.

Patient Characteristics Associated with Telehealth Scheduling and Completion in Primary Care at a Large, Urban Public Healthcare System.

Understanding patient characteristics associated with scheduling and completing telehealth visits can identify potential biases or latent preferences related to telehealth usage. We describe patient characteristics associated with being scheduled for and completing audio and video visits. We used data from patients at 17 adult primary care departments in a large, urban public healthcare system from August 1, 2020 to July 31, 2021. We used hierarchical multivariable logistic regression to generate adjusted odds ratios (aOR) for patient characteristics associated with having been scheduled for and completed telehealth (vs in-person) visits and for video (vs audio) scheduling and completion during two time periods: a telehealth transition period (N = 190,949) and a telehealth elective period (N = 181,808). Patient characteristics were significantly associated with scheduling and completion of telehealth visits. Many associations were similar across time periods, but others changed over time. Patients who were older (≥ 65years old vs 18-44years old: aOR for scheduling 0.53/completion 0.48), Black (0.86/0.71), Hispanic (0.76/0.62), or had Medicaid (0.93/0.84) were among those less likely to be scheduled for or complete video (vs audio) visits. Patients with activated patient portals (1.97/3.34) or more visits (≥ 3 scheduled visits vs 1 visit: 2.40/1.52) were more likely to be scheduled for or complete video visits. Variation in scheduling/completion explained by patient characteristics was 7.2%/7.5%, clustering by provider 37.2%/34.9%, and clustering by facility 43.1%/37.4%. Stable and dynamic associations suggest persistent gaps in access and evolving preferences/biases. Variation explained by patient characteristics was relatively low compared with that explained by provider and facility clustering.

News recommendations based on collaborative topic modeling and collaborative filtering with generative adversarial networks

PurposeOnline news websites provide huge amounts of timely news, bringing the challenge of recommending personalized news articles. Generative adversarial networks (GAN) based on collaborative filtering (CFGAN) can achieve effective recommendation quality. However, CFGAN ignores item contents, which contain more latent preference features than just user ratings. It is important to consider both ratings and item contents in making preference predictions. This study aims to improve news recommendation by proposing a GAN-based news recommendation model considering both ratings (implicit feedback) and the latent features of news content.Design/methodology/approachThe collaborative topic modeling (CTM) can improve user preference prediction by combining matrix factorization (MF) with latent topics of item content derived from latent topic modeling. This study proposes a novel hybrid news recommendation model, Hybrid-CFGAN, which modifies the architecture of the CFGAN model with enhanced preference learning from the CTM. The proposed Hybrid-CFGAN model contains parallel neural networks – original rating-based preference learning and CTM-based preference learning, which consider both ratings and news content with user preferences derived from the CTM model. A tunable parameter is used to adjust the weights of the two preference learnings, while concatenating the preference outputs of the two parallel neural networks.FindingsThis study uses the dataset collected from an online news website, NiusNews, to conduct an experimental evaluation. The results show that the proposed Hybrid-CFGAN model can achieve better performance than the state-of-the-art GAN-based recommendation methods. The proposed novel Hybrid-CFGAN model can enhance existing GAN-based recommendation and increase the performance of preference predictions on textual content such as news articles.Originality/valueAs the existing CFGAN model does not consider content information and solely relies on history logs, it may not be effective in recommending news articles. Our proposed Hybrid-CFGAN model modified the architecture of the CFGAN generator by adding a parallel neural network to gain the relevant information from news content and user preferences derived from the CTM model. The novel idea of adjusting the preference learning from two parallel neural networks – original rating-based preference learning and CTM-based preference learning – contributes to improve the recommendation quality of the proposed model by considering both ratings and latent preferences derived from item contents. The proposed novel recommendation model can improve news recommendation, thereby increasing the commercial value of news media platforms.

Talent recommendation based on attentive deep neural network and implicit relationships of resumes

Talent recruitment has become a crucial issue for companies since finding suitable candidates from the massive data on potential candidates from online talent platforms is a challenging task. However, extant studies mainly focus on the scalability and inference ability of models, while the dynamic variability of the importance of each feature in different scenarios is barely addressed. Besides, there is a lack of research on how to depict the hidden potential preference of the job which cannot be derived from job requirements. In this paper, we propose a two-stage resume recommendation model based on deep learning and attention mechanisms, especially considering the latent preference information hidden in the hired employee resumes, named the Attentive Implicit Relationship-Aware Neural Network (AIRANN) model. Specifically, a novel mechanism is proposed herein to extract the hidden potential preference of the corresponding job by deriving the implicit relationship between the target resume and hired employees’ resumes. Existing studies have not considered such resume implicit relationships. Moreover, we propose a Feature Co-Attention mechanism to capture the dynamic interactive importance within the non-text features of both resumes and jobs. For different jobs, the suitability of resumes would be valued from different aspects, including resume implicit relationships, as well as textual and non-textual features. Accordingly, an Aspect-attention mechanism is designed herein to automatically adjust the variant importance of each aspect. Finally, extensive experiments are conducted on a real-world company dataset. The experiment results of ablation studies demonstrate the effectiveness of each mechanism in the proposed AIRANN model. The experiment results also show that the proposed AIRANN model outperforms other baseline methods, showing a general improvement of 13.31%, 12.49%, 6.5% and 7.17% over the state-of-the-art baseline under F1@6, F1@15, NDCG@6 and NDCG@15, respectively.

Access to social welfare and migrants’ intercity location choices in the case of China: Do latent preferences matter?

ABSTRACT Access to social welfare has been downplayed in understanding migrants’ decisions about destinations and settlements. Moreover, how migrants’ latent mobility preferences affect their preferences for migration drivers remains underexplored. This study uses latent class logit models to identify three types of internal migrants with distinctive latent mobility preferences in China. The findings suggest that access to social welfare is the most important predictor of decisions about destination cities for all types of migrants, particularly those with low mobility preference. For highly mobile migrants, access to social welfare is an important consideration but expected wages and urban amenities also play important roles in their location choices and settlement. The findings highlight the crucial role of social welfare in the relocation and settlement of internal migrants in China. The heterogeneous preferences regarding migration drivers among migrants with different latent mobility preferences also call for tailor-made local labor policymaking.