Search results

This study aims to provide the best estimate of a stock's next day's closing price for a given day with the help of the hidden Markov model–Gaussian mixture model (HMM-GMM). The results were compared with Hassan and Nath’s (2005) study using HMM and artificial neural network (ANN).

The study adopted an initialization approach wherein the hidden states of the HMM are modelled as GMM using two different approaches. Training of the HMM-GMM model is carried out using two methods. The prediction was performed by taking the closest closing price (having a log-likelihood within the tolerance range) to that of the present one as the closing price for the next day. Mean absolute percentage error (MAPE) has been used to compare the proposed GMM-HMM model against the models of the research study (Hassan and Nath, 2005).

Comparing this study with Hassan and Nath (2005) reveals that the proposed model outperformed in 66 out of the 72 different test cases. The results affirm that the model can be used for more accurate time series prediction. Further, compared with the results of the ANN model from Hassan's study, the proposed HMM model outperformed 24 of the 36 test cases.

The study introduced a novel initialization and two training/prediction approaches for the HMM-GMM model. It is to be noted that the study has introduced a GMM-HMM-based closing price estimator for stock price prediction. The proposed method of forecasting the stock prices using GMM-HMM is explainable and has a solid statistical foundation.

This paper aims to present a method based on hidden Markov models (HMM) for extracting information from web news.

The samples under study are derived from the contents of PROC “People's Daily Online,” a web‐based news publication containing non‐structured archives. This study focuses on developing HMM‐based tools for news filtering in order to retrieve terms of interest, such as “Geo‐location,” “System,” and “Personas.” The experiments are performed in two stages. In the first stage, each HMM being built is exclusively serving for extracting unique target term in order to evaluate the fundamental information extraction (IE) capability. In the second stage, the experiment is then extended to resolve a more complex, multi‐term extraction issue.

The results reveal that, by using HMMs as a basis, the accuracies (F‐measure) for unique IE tasks can achieve more than 70 per cent on average, while no fewer than 66 per cent accuracies are obtained for multi‐term extraction.

The study reveals the promising of using HMM for developing automatic tool in filtering free‐structured data.

The purpose of this paper is to propose an automatic metadata extraction and retrieval system to extract bibliographical information from digital academic documents in portable document formats (PDFs).

The authors use PDFBox to extract text and font size information, a rule-based method to identify titles, and an Hidden Markov Model (HMM) to extract the titles and authors. Finally, the extracted titles and authors (possibly incorrect or incomplete) are sent as query strings to digital libraries (e.g. ACM, IEEE, CiteSeerX, SDOS, and Google Scholar) to retrieve the rest of metadata.

Four experiments are conducted to examine the feasibility of the proposed system. The first experiment compares two different HMM models: multi-state model and one state model (the proposed model). The result shows that one state model can have a comparable performance with multi-state model, but is more suitable to deal with real-world unknown states. The second experiment shows that our proposed model (without the aid of online query) can achieve as good performance as other researcher's model on Cora paper header dataset. In the third experiment the paper examines the performance of our system on a small dataset of 43 real PDF research papers. The result shows that our proposed system (with online query) can perform pretty well on bibliographical data extraction and even outperform the free citation management tool Zotero 3.0. Finally, the paper conducts the fourth experiment with a larger dataset of 103 papers to compare our system with Zotero 4.0. The result shows that our system significantly outperforms Zotero 4.0. The feasibility of the proposed model is thus justified.

For academic implication, the system is unique in two folds: first, the system only uses Cora header set for HMM training, without using other tagged datasets or gazetteers resources, which means the system is light and scalable. Second, the system is workable and can be applied to extracting metadata of real-world PDF files. The extracted bibliographical data can then be imported into citation software such as endnote or refworks to increase researchers’ productivity.

For practical implication, the system can outperform the existing tool, Zotero v4.0. This provides practitioners good chances to develop similar products in real applications; though it might require some knowledge about HMM implementation.

The HMM implementation is not novel. What is innovative is that it actually combines two HMM models. The main model is adapted from Freitag and Mccallum (1999) and the authors add word features of the Nymble HMM (Bikel et al, 1997) to it. The system is workable even without manually tagging the datasets before training the model (the authors just use cora dataset to train and test on real-world PDF papers), as this is significantly different from what other works have done so far. The experimental results have shown sufficient evidence about the feasibility of our proposed method in this aspect.

Cognitive radio (CR) plays a very important role in enabling spectral efficiency in wireless communication networks, where the secondary user (SU) allows the licensed primary users (PUs). The purpose of this paper is to develop a prediction model for spectrum sensing in CR.

This paper proposes a hybrid prediction model, called krill-herd whale optimization-based actor critic neural network and hidden Markov model (KHWO-ACNN-HMM). The spectral bands are determined optimally using the proposed hybrid prediction model for allocating the spectrum bands to the PUs. For better sensing, the eigenvalue based on cooperative sensing used in CR. Finally, a hybrid model is designed by hybridizing KHWO-ACNN and HMM to enhance the accuracy of sensing. The predicted results of KHWO-ACNN and HMM are combined by a fusion model, for which a weighted entropy fusion is employed to determine the free spectrum available in CRs.

The performance of the prediction model is evaluated based on metrics, such as probability of detection, probability of false alarm, throughput and sensing time. The proposed spectrum sensing method achieves maximum probability of detection of 0.9696, minimum probability of false alarm rate as 0.78, minimum throughput of 0.0303 and the maximum sensing time of 650.08 s.

The proposed method is useful in various applications, including authentication applications, wireless medical networks and so on.

A hybrid prediction model is introduced for energy efficient spectrum sensing in CR and the performance of the proposed model is evaluated with the existing models. The proposed hybrid model outperformed the other techniques.

The paper aims to propose a novel strategic approach, named a Scorecard‐Markov model, combining an evaluation scorecard and a hidden Markov model (HMM) for new product idea screening (NPIS) decisions.

A scorecard is constructed to evaluate new product ideas on several criteria, including customer needs, marketing strength, competency, manufacturing compatibility, and distribution channels, involving a consideration of risk buy. A HMM is then developed accordingly to predict the overall performance of new ideas in terms of success probability. To implement the model, it is trained and tested by the historical dataset of a world‐class, leading company in the power tools industry through a case study.

The approach is proven to be encouraging and meaningful. The scorecard can serve as a guide for new product idea evaluation to convert experts' linguistic judgments to quantifiable and comparable data, whereas the HMM can determine the success probability of new product ideas to support NPIS decision making based on their computed evaluation performance. The optimal cut‐off value for making either a go or kill decision on each idea can thus be determined. Concerning the case company, a go decision should be made when the probability lies in the interval [0.53, 1].

The model can prevent companies from undertaking risky and failed new product development projects. Further, it is believed that this study can assist decision makers in choosing winning new product ideas towards commercialization in an effective and certain manner, thus enhancing the new product success rate in the innovation industry.

The approach incorporating the scorecard method and HMM is novel. Illustrated by the case study, the application of this approach to NPIS decisions is confirmed to be effective.

The aim of this paper is to develop autonomous mobile home healthcare robots, which are capable of observing patients’ motions, recognizing the patients’ behaviours based on observation data, and providing automatically calling for medical personnel in emergency situations. The robots to be developed will bring about cost‐effective, safe and easier at‐home rehabilitation to most motor‐function impaired patients (MIPs).

The paper has developed following programs/control algorithms: control algorithms for a mobile robot to track and follow human motions, to measure human joint trajectories, and to calculate angles of lower limb joints; and algorithms for recognizing human gait behaviours based on the calculated joints angle data.

A Hidden Markov Model (HMM) based human gait behaviour recognition taking lower limb joint angles and body angle as input was proposed. The proposed HMM based gait behaviour recognition is compared with the Nearest Neighbour (NN) classification methods. Experimental results showed that a human gait behaviour recognition using HMM can be achieved from the lower limb joint trajectory with higher accuracy than compared classification methods.

The research addresses human motion tracking and recognition by a mobile robot. Human gait behaviour recognition is HMM based lower limb joints and body angle data from extracted from kinect sensor at the mobile robot.

By using a new feature extraction method on the Cert data set and using a hidden Markov model (HMM) to model and analyze the behavior of users to distinguish whether the behavior is normal within a continuous period.

Feature extraction of five parts of the time series by rules and sorting in chronological order. Use the obtained features to calculate the probability parameters required by the HMM model and establish a behavior model for each user. When the user has abnormal behavior, the model will return a very low probability value to distinguish between normal and abnormal information.

Generally, HMM parameters are obtained by supervised learning and unsupervised learning, but the hidden state cannot be clearly defined. When the hidden state is determined according to the data set, the accuracy of the model will be improved.

This paper proposes a new feature extraction method and analysis mode, which determines the shape of the hidden state according to the situation of the data set, making subsequent HMM modeling simple and efficient and in turn improving the accuracy of user behavior detection.

Behavioural change practice has focussed attention on understanding behaviour; failing to apply dynamic approaches that capture the underlying determinants of behavioural change. Following recommendations to direct analytical focus towards understanding both the causal factors of behaviour and behavioural change to enhance intervention practice, this paper aims to apply a hidden Markov model (HMM) approach to understand why people transition from one state to another (e.g. reporting changes from wasting food to not wasting food or vice versa).

Data were drawn from a 2017 food waste programme that aimed to reduce waste of fruit and vegetables by increasing self-efficacy through a two-week pilot, featuring recipes and in-store cooking demonstrations. A repeated measure longitudinal research design was used. In total, 314 households completed a phone survey prior to the two-week pilot and 244 completed the survey in the weeks following the intervention (77% retention in the evaluation study).

Two behavioural states were identified, namely, fruit and vegetable (FV) wasters and non-FV wasters. Age was identified as a causal factor for FV food wasting prior to the campaign (45-54 years were most likely to waste FV). Following the intervention, a total of 43.8% transitioned away from FV wasters to non-wasters, and attitudes and self-efficacy were indicated as potential causal factors of this change in FV waste behaviour.

Through this application, it is demonstrated how HMM can identify behavioural states, rates of behaviour change and importantly how HMM can identify both causal determinants of behaviour and behavioural change. Implications, limitations and future research directions are outlined.

This paper aims to demonstrate how consumer clickstream data from a leading hotel search engine can be used to validate two hidden information processing stages – first eliminate alternatives, then choose – proposed by the revered information processing theory of consumer choice.

This study models the two hidden information processing stages as hidden states in a hidden Markov model, estimated on consumer search behavior, product attributes and diversity of alternatives in the consideration set.

First, the stage of information processing can be statistically characterized in terms of consumer search covariates, including trip characteristics, use of search tools and the diversity of the consideration set, operationalized in terms of: number of brands, dispersion of price and dispersion of quality. Second, users are more sensitive to price and quality in the first rather than the second stage, which is closer to purchase.

The results suggest practical implications for how search engine managers can target consumers with appropriate marketing-mix actions, based on which information processing stage consumers might be in.

Most previous studies on validating the information processing theory of consumer choice have used laboratory experiments, subjects and information display boards comprising hypothetical product alternatives and attributes. Only a few studies use observational data. In contrast, this study uniquely uses point-of-purchase clickstream data on actual visitors at a leading hotel search engine and tests the theory based on real products, attributes and diversity of the consideration set.

To delineate the range of consumer responses to life‐conditions where sustained powerlessness is experienced. To provide a framework to understand the ways in which these consumers try to reclaim degrees of self‐empowerment and wellbeing.

Goffman's conceptualization of stigma is employed to study a heavy metal music enclave consisting of lower socioeconomic consumers, who exhibit a range of stigmatizing attributes.

A taxonomy of ten consumer remedies for their situation is developed. These include: resignation, confrontation, withdrawal, engagement, concealment, escapism, hedonic, spiritual, nostalgia, and creative. Each can potentially have negative or positive consequences. However, we found consumers often use a blend of these remedies as pathways to self‐empower.

Understanding the strengths and weaknesses of each of the remedies will potentially guide public policy makers in shaping programs better able to foster self‐empowerment among disadvantaged consumers.

The paper advances understanding of consumer response to sustained powerlessness as consequence of disadvantaged life conditions that are resistant to change.