Ensemble machine learning for travel time prediction: An Active learning-assisted approach

Travel Time Prediction (TTP) has become an essential service that people use in daily commutes. With the precise TTP, individuals, logistic companies, and transport authorities can better manage their activities and operations. This paper presents a novel Hybridized Deep Feature Space (HDFS) based TTP ensemble model (HDFS-TTP) for accurate travel time prediction. In the first step, extensive endogenous and exogenous data sources are augmented with traffic data obtained using sensors. Next, we used Principal Component Analysis (PCA) and Deep Stacked Auto-Encoder (DSAE) for feature reduction. We generated feature spaces of deep learning models, namely Convolutional Neural Network (CNN), Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU), and fed them to a model based on Support Vector Regressor (SVR) for predicting travel times. Two best-performing models are selected, and their feature spaces are hybridized to boost feature space. On this boosted feature space, we employed SVR for final prediction. Our proposed HDFS-TTP ensemble can learn complex non-linearities in traffic data with the varying architectural design. The performance of our proposed HDFS-TTP ensemble using hybridized and boosted feature spaces showed significant improvement in test data in terms of Root Mean Square Error (62.27± 1.58), Mean Absolute Error (13.38± 1.09), Maximum Absolute Error (104.66± 2.77), Mean Absolute Percentage Error (2.50± 0.03), and Coefficient of determination (0.99714± 0.00044).

Recently, the amount of trajectory data has been rapidly increasing with the popularity of LBS and the development of mobile technology. Thus, the analysis of trajectory patterns for large amounts of trajectory data has attracted much interest. To improve the quality of trajectory-based services, it is essential to predict an exact travel time for a given query on road networks. One of the typical schemes for travel time prediction is a rule-based classification method which can ensure high accuracy. However, the existing scheme is inadequate for the processing of massive data because it is designed without the consideration of distributed computing environments. To solve this problem, this paper proposes a travel time prediction algorithm using rule-based classification on MapReduce for a large amount of trajectory data. First, our algorithm generates classification rules based on the actual traffic statistics and measures adequate velocity classes for each road segment. Second, our algorithm generates a distributed index by using the grid-based map partitioning method. Our algorithm can reduces the query processing cost because it only retrieves the grid cells which contain a query region, instead of the entire road network. Furthermore, it can reduce the query processing time by estimating the travel time for each segment of a given query in a parallel way. Finally, we show from our performance analysis that our scheme performs more accurate travel time prediction than the existing algorithms.

Forecasting journey time distribution with consideration to abnormal traffic conditions

The world is changing so fast, and therefore education needs to adapt to the challenges of times. In order to update the content of school education in the Republic of Kazakhstan modern trends are going to be used. These trends contain pedagogical methods that can be used to preserve and even increase internal motivation, as active learning. Active learning method is an treatment where students participate or interact with the learning process, as opposed to passively taking in the information.The goal of this study is to identify the impact of active learning method on 10th grade students’ attitude towards mathematics of the students the second semester of the school year 2019-2020. More specifically, it attempted to determine and compare the attitude toward mathematics of students’ exposure to active learning and traditional teaching strategy. The Likert scale used to evaluate the attitude of students toward mathematics. Mean, Cronbach value, T-test were the statistical tools used in anatomizing and interpreting the research data. The discovering showed that the students in the active learning group had auspicious attitude than students in the conventional teaching group. According to the findings after research, we saw the direct relation between attitude and active learning. It is concluded that the students’ attitude toward mathematics was better by using active learning strategy. It is recommended that mathematics teacher should use active learning strategy in order to improve the attitude toward mathematics of the students.Keywords: attitude, mathematics, active learning

Active learning is one form of supervised machine learning. In supervised learning, a set of labeled samples is passed to a learning algorithm for training a classifier. However, labeling large amounts of training samples can be costly and error-prone. Active learning deals with the development of algorithms that interactively select a subset of the available unlabeled samples for labeling, and aims at minimizing the labeling effort while maintaining classification performance. The key challenge for the development of so-called active learning strategies is the balance between exploitation and exploration: On the one hand, the estimated decision boundary needs to be refined in feature space regions where it has already been established, while, on the other hand, the feature space needs to be scanned carefully for unexpected class distributions. In this thesis, two approaches to active learning are presented that consider these two aspects in a novel way. In order to lay the foundations for the first one, it is proposed to express the uncertainty in class prediction of a classifier at a test point in terms of a second-order distribution. The mean of this distribution corresponds to the common estimate of the posterior class probabilities and thus is related to the distance of the test point to the decision boundary, whereas the spread of the distribution indicates the degree of exploration in the corresponding region of feature space. This allows for the evaluation of the utility of labeling a yet unlabeled point with respect to classifier improvement in a principled way and leads to a completely novel approach to active learning. The proposed strategy is then implemented and evaluated based on kernel density classification. The generic active learning strategy can be combined with any other classifier, but it performs best if the derived second-order distributions are sufficiently good approximations to the sampling distribution. Although second-order distributions for random forests are derived in this thesis, they do not approximate sufficiently well the sampling distribution and mainly allow only for the relative comparison of prediction uncertainty between test points. In order to combine the state of the art classification performance of random forests with the principal ideas of the first active learning approach, a related second approach for random forests is derived. It is, in addition, tailored to the demands in industrial optical inspection: bag-wise labeling with weak labels and strongly imbalanced classes. Moreover, an outlier detection scheme based on random forests is derived that is used by the proposed active learning algorithm. Finally, a new computational scheme for Gaussian process classification is presented. It is compared to two standard methods in geostatistics, both with respect to theoretical consistency and practical performance. The method evolved as a by-product when considering using Gaussian process models for active learning.

Therapists’ active learning increases treatment fidelity, but research is needed on supervisory strategies to engage therapists in active learning. This study used sequential analysis to examine consultant behaviors associated with increased and decreased probability of eliciting therapists’ active learning. The study included 162 consultation sessions from 27 community therapists implementing Attachment and Biobehavioral Catch-up. Consultants’ client discussion, information provision, and modeling were associated with reduced likelihood of active learning. Consultants’ questions, engagement in active learning strategies, use of video, and silence were associated with greater likelihood of therapist active learning. These findings inform supervisors’ attempts to encourage active learning.

Deep learning has recently attracted significant attention in the field of hyperspectral images (HSIs) classification. However, the construction of an efficient deep neural network (DNN) mostly relies on a large number of labeled samples being available. To address this problem, this paper proposes a unified deep network, combined with active transfer learning that can be well-trained for HSIs classification using only minimally labeled training data. More specifically, deep joint spectral-spatial feature is first extracted through hierarchical stacked sparse autoencoder (SSAE) networks. Active transfer learning is then exploited to transfer the pre-trained SSAE network and the limited training samples from the source domain to the target domain, where the SSAE network is subsequently fine-tuned using the limited labeled samples selected from both source and target domain by corresponding active learning strategies. The advantages of our proposed method are threefold: 1) the network can be effectively trained using only limited labeled samples with the help of novel active learning strategies; 2) the network is flexible and scalable enough to function across various transfer situations, including cross-dataset and intra-image; 3) the learned deep joint spectral-spatial feature representation is more generic and robust than many joint spectral-spatial feature representation. Extensive comparative evaluations demonstrate that our proposed method significantly outperforms many state-of-the-art approaches, including both traditional and deep network-based methods, on three popular datasets.

Faculty motivations to use active learning among pharmacy educators

Molecular dynamics with predefined functional forms is a popular technique for understanding dynamical evolution of systems. The predefined functional forms impose limits on the physics that can be captured. Artificial neural network (ANN) models have emerged as an attractive flexible alternative to the expensive quantum calculations (e.g., density functional theory) in the area of molecular force-fields. Ideally, if one is able to train a ANN to accurately predict the correct DFT energy and forces for any given structure, they gain the ability to perform molecular dynamics with high accuracy while simultaneously reducing the computation cost in a dramatic fashion. While this goal is very lucrative, neural networks are interpolative and therefore, it is not always clear how one should go about training a neural network to exhaustively fit the entire phase space of a given system. Currently, ANNs are trained by generating large quantities (on the order of 104 or greater) of training data in hopes that the ANN has adequately sampled the energy landscape both near and far-from-equilibrium. This can, however, be a bit prohibitive when it comes to more accurate levels of quantum theory. As such, it is desirable to train a model using the absolute minimal data set possible, especially when costs of high-fidelity calculations such as CCSD and QMC are high. Here, we present an active learning approach that iteratively trains an ANN model to faithfully replicate the coarse-grained energy surface of water clusters using only 426 total structures in its training data. Our active learning workflow starts with a sparse training data set which is continually updated via a Nested Ensemble Monte Carlo scheme that sparsely queries the energy landscape and tests the network performance. Next, the network is retrained with an updated training set that includes failed configurations/energies from previous iteration until convergence is attained. Once trained, we generate an extensive test set of 100 000 configurations sampled across clusters ranging from 1 to 200 molecules and demonstrate that the trained network adequately reproduces the energies (within mean absolute error (MAE) of 2 meV/molecule) and forces (MAE 40 meV/Å) compared to the reference model. More importantly, the trained ANN model also accurately captures both the structure as well as the free energy as a function of the various cluster sizes. Overall, this study reports a new active learning scheme with promising strategy to develop accurate force-fields for molecular simulations using extremely sparse training data sets.

본 연구는 Deep Neural Network(DNN)을 이용하여 광주-기아 챔피언스 필드의 일일 관중 수를 예측함으로써 이를 통해 구단과 관련기업의 마케팅 자료제공 및 구장 내 부대시설의 재고관리에 자료로 쓰임을 목적으로 수행 되었다. 본 연구 에서는 Artificial Neural Network(ANN)의 종류인 DNN 모델을 이용하였으며 DNN 모델의 과적합을 막기 위해 Dropout 과 Batch normalization 적용한 모델을 바탕으로 총 4종류를 설계하였다. 각각 10개의 DNN을 만들어 예측값의 Root Mean Square Error(RMSE)와 Mean Absolute Percentage Error(MAPE)의 평균값을 낸 모델과 예측값의 평균으로 RMSE와 MAPE를 평가한 Ensemble 모델을 만들었다. 모델의 학습 데이터는 2008년부터 2017년까지의 관중 수 데이터를 수집하여 수집된 데이터의 80%를 무작위로 선정하였으며, 나머지 20%는 테스트 데이터로 사용하였다. 총 100회의 데이터 선정, 모델구성 그리고 학습 및 예측을 한 결과 Ensemble 모델은 DNN 모델의 예측력이 가장 우수하게 나왔으며, 다중선 형회귀 모델 대비 RMSE는 15.17%, MAPE는 14.34% 높은 예측력을 보이고 있다.In this paper, we used the Deep Neural Network (DNN) to predict the number of daily spectators of Gwangju -KIA Champions Field in order to provide marketing data for the team and related businesses and for managing the inventories of the facilities in the stadium. In this study, the DNN model, which is based on an artificial neural network (ANN), was used, and four kinds of DNN model were designed along with dropout and batch normalization model to prevent overfitting. Each of four models consists of 10 DNNs, and we added extra models with ensemble model. Each model was evaluated by Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). The learning data from the model randomly selected 80% of the collected data from 2008 to 2017, and the other 20% were used as test data. With the result of 100 data selection, model configuration, and learning and prediction, we concluded that the predictive power of the DNN model with ensemble model is the best, and RMSE and MAPE are 15.17% and 14.34% higher, correspondingly, than the prediction value of the multiple linear regression model.

Aim/Purpose: Two popular methods for encouraging active learning are Augmented Reality (AR) and Virtual Field Trips (VFTs). This exploratory case study examines college students’ perceptions of a prototype AR and VFT app as an active learning strategy. Background: AR allows students to learn as they physically explore a destination, while VFTs give students the opportunity to visit exciting destinations without leaving their homes. AR and VFTs promote active learning, which has been shown to increase college student success in Science, Technology, Engineering, and Math (STEM) courses. The aim of the VFT app in this study is to provide college students in a STEM course with an interactive lesson on modeling information systems using diagrams. Methodology: This exploratory case study is intended to serve as a condensed case study performed with the prototype version of a VFT app before implementing a large-scale investigation of students’ perceptions of a more refined version of the app. The study employed a qualitative approach involving a survey with open-ended questions to gather college students’ perceptions of learning with a VFT. The data were analyzed using inductive coding. The participants are students at a mid-sized, urban, public university. Contribution: This exploratory case study serves as a proof-of-concept and starting point for other faculty who may be interested in developing their own AR and VFT apps to engage students in active learning. Releasing the app to a common Open Educational Resources (OER) repository will give other faculty easy access to re-use the app and build upon it to create their own virtual field trips. OER are learning materials that are freely available for students and faculty to download and use in their coursework. Findings: Students overwhelmingly perceived the VFT app helped them learn about the subject that was presented, citing the visual nature of the app, the real-life scenarios presented in the app, and the app’s ease of use as reasons why. The majority (over 89%) also agreed that the app motivated them to learn more about the subject, mainly due to the app’s real-life scenarios, and over 83% of students cited at least one benefit to learning with the app, such as the navigation/location features, the easy-to-use interface, and the real-world scenarios. Recommendations for Practitioners: The pedagogical implications of this study are that faculty should adopt VFTs as an active learning strategy, particularly in STEM college courses, based on the students’ positive perceptions of learning, motivation, and benefits of VFTs. Recommendations for Researchers: Researchers can expand on this exploratory case study by conducting a larger-scale study of the VFT app employed in the case study, or by developing their own VFT app based on the one in this study, to capture a broader group of students’ perceptions of VFTs as an active learning strategy. Impact on Society: The broad impact of this research on society is encouraging the adoption of VFTs as an active learning strategy since active learning strategies are shown to increase college students’ success and engagement. Future Research: Future research will be conducted in subsequent terms to gather additional data on students’ perceptions of the VFT app, as well as their perceptions of the relationship between learning and the VFT technology. Further research is also needed to survey faculty on their perceptions of how engaging with the app impacts student learning, particularly in regards to the VFT technology within the app.

Traffic states prediction in air transportation systems is a challenging problem and has not been fully explored because it is subject to many more highly correlated factors and a more complicated traffic management scheme compared to urban transportation systems. It becomes a more formidable task when facing a multi-airport system (MAS), in which several major airports are closely located and tightly coupled with each other through limited terminal airspace. In this work, we propose a novel method using a time series model and recurrent neural network to make the estimated time of arrival (ETA) for a flight to an MAS, which can be potentially utilized for flight delay prediction and congestion analysis. The experiment utilizes two months of 4D trajectories data from Beijing Capital International Airport (PEK) to Shenzhen Bao’an International airport (ZGSZ). The entire prediction work is decomposed into two sub-problems, en-route travel time prediction which is from flight origin to the entering gate of MAS, defined as the location is 300km from the airport in MAS, and terminal maneuvering area (TMA) travel time prediction which is from the entrance to flight’s destination. The auto-regressive integrated moving average (ARIMA), a time series prediction model, is used to predict travel time in en-route under given the flight departure time. Bidirectional long short term memory (LSTM), a recurrent neural network, is developed to forecast travel time in the arrival approach by utilizing spatio-temporal features. To design the input features, we use density-based spatial clustering (DBSCAN) with the help of the Voronoi diagram to extract spatial information from every historical flight trajectory of aircraft operated in an MAS, then select the observation time window to capture the temporal information for each flight. The Multivariate Stacked Fully connected-Bidirectional LSTM (MSFCB-LSTM) model is constructed to make shortterm forecasting using spatio-temporal features we designed when the flight’s entering MAS time is given. For TMA travel time prediction, a case study of Guangdong-Hong Kong-Macao Greater Bay Area (GHM-GBA), a typical MAS which contains five major airports closely located within 120km, is carried out using actual historical 4D trajectory data. Finally, Using two months 4D trajectories data, PEK to ZGSZ, the result exhibits the best accuracy, a measurement we define for prediction, of the longterm prediction of ETA given departure time is 92%, and mean absolute error (MAE) is 6.09 minutes.

Existing meta-analyses have shown that active learning strategies are effective in improving students’ learning performance. However, their implementation may vary across school levels. This study investigated differences among elementary, middle, and high school teachers in the use of teaching strategies, including those that promote active learning and pedagogical beliefs. We also explored differences in the relationships between beliefs about teaching and learning, teaching experience, and teaching strategies for active learning across school levels. An online survey was conducted with 550 in-service elementary, middle, and high school teachers. Participants completed a questionnaire that measured their teaching strategies, beliefs about teaching and learning, and teaching experience. The results revealed differences in the frequency of use of six teaching strategies, including those that promoted active learning. Elementary school teachers used teaching strategies that promoted active learning most frequently, followed by middle and high school teachers. There were no differences in teachers’ beliefs about teaching and learning across school levels. Multigroup structural equation modeling indicated no differences in the influence of beliefs about teaching and learning on active learning teaching strategies among school levels, except that the traditional conception was negatively associated with the implementation of active learning in middle school. Constructivist beliefs were positively associated with active learning strategies across all levels, whereas teaching experience was negatively associated with teaching strategies that promoted student output at the higher school level. These findings have implications for the implementation of active learning, particularly at higher school levels.

Modern Information and Communication Technology (ICT)-based applications utilize current technological advancements for purposes of streaming data, as a way of adapting to the ever-changing technological landscape. Such efforts require providing accurate, meaningful, and trustworthy output from the streaming sensors particularly during dynamic virtual sensing. However, to ensure that the sensing ecosystem is devoid of any sensor threats or active attacks, it is paramount to implement secure real-time strategies. Fundamentally, real-time detection of adversarial attacks/instances during the User Feedback Process (UFP) is the key to forecasting potential attacks in active learning. Also, according to existing literature, there lacks a comprehensive study that has a focus on adversarial detection from an active machine learning perspective at the time of writing this paper. Therefore, the authors posit the importance of detecting adversarial attacks in active learning strategy. Attack in the context of this paper through a UFP-Threat driven model has been presented as any action that exerts an alteration to the learning system or data. To achieve this, the study employed ambient data collected from a smart environment human activity recognition from (Continuous Ambient Sensors Dataset, CASA) with fully labeled connections, where we intentionally subject the Dataset to wrong labels as a targeted/manipulative attack (by a malevolent labeler) in the UFP, with an assumption that the user-labels were connected to unique identities. While the dataset's focus is to classify tasks and predict activities, our study gives a focus on active adversarial strategies from an information security point of view. Furthermore, the strategies for modeling threats have been presented using the Meta Attack Language (MAL) compiler for purposes adversarial detection. The findings from the experiments conducted have shown that real-time adversarial identification and profiling during the UFP could significantly increase the accuracy during the learning process with a high degree of certainty and paves the way towards an automated adversarial detection and profiling approaches on the Internet of Cognitive Things (ICoT).

Although the relationship between student motivation, self-perception, and achievement in science education is well established, there is still a lack of focused studies on these aspects at the lower secondary level, particularly in private school settings where student demographics and learning environments may pose unique challenges. The purpose of this study is to: (1) assess lower secondary students' motivation and self-concept in scientific learning; and (2), using correlational analysis and qualitative feedback, determine ways to improve these aspects. Using a quantitative descriptive survey design, this study assessed 60 grade 7-9 students from a private lower secondary school. The Students’ Motivation Toward Science Learning (SMTSL) and adapted Chemistry Self-Concept Inventory (CSCI) measured motivation and self-concept, respectively. Data analysis included mean scores for both variables and Spearman’s rank correlation (accounting for Liker-scale ordinality and non-normal distributions). A moderate-high level of motivation (Active learning strategies [ALS]=4.0-4.1 highest, Performance Goal [PG]=2.8-3.1 lowest) and self-concept (Positive Perception of Science Self Concept [PSSC]=3.8-3.9 highest, Mathematic Self-Concept [MSC]=2.9-3.2 lowest) with moderate positive correlation (r = 0.572, p < 0.001) underscores that self-concept and motivation are interrelated, suggests interventions should simultaneously promote active learning (to enhance ALS) while addressing mathematical confidence (to improve MSC). These findings suggest that teachers can boost engagement through active learning and confidence-building strategies, while school and policymakers should design targete interventions to support students with low scientific confidence and high performance pressure.