Past Sessions Research Articles

Evolving intra-and inter-session graph fusion for next item recommendation

Next-item recommendation aims to predict users’ subsequent behaviors using their historical sequence data. However, sessions are often anonymous, short, and time-varying, making it challenging to capture accurate and evolving item representations. Existing methods using static graphs may fail to model the evolving semantics of items over time. To address this problem, we propose the Evolving Intra-session and Inter-session Graph Neural Network (EII-GNN) to capture the evolving item semantics by fusing global and local graph information. EII-GNN utilizes a global dynamic graph to model inter-session item transitions and update item embeddings at each timestamp. It also constructs a per-session graph with shortcut edges to learn complex intra-session patterns. To personalize recommendations, a history-aware GRU applies the user’s past sessions. We fuse the inter-session graph, intra-session graph, and history embeddings to obtain the session representation for final recommendation. Our model performed well in experiments with three real-world data sets against its state-of-the-art counterparts.

How to reward animals based on their subjective percepts: A Bayesian approach to online estimation of perceptual biases.

Elucidating the neural basis of perceptual biases, such as those produced by visual illusions, can provide powerful insights into the neural mechanisms of perceptual inference. However, studying the subjective percepts of animals poses a fundamental challenge: unlike human participants, animals cannot be verbally instructed to report what they see, hear, or feel. Instead, they must be trained to perform a task for reward, and researchers must infer from their responses what the animal perceived. However, animals' responses are shaped by reward feedback, thus raising the major concern that the reward regimen may alter the animal's decision strategy or even intrinsic perceptual biases. We developed a method that estimates perceptual bias during task performance and then computes the reward for each trial based on the evolving estimate of the animal's perceptual bias. Our approach makes use of multiple stimulus contexts to dissociate perceptual biases from decision-related biases. Starting with an informative prior, our Bayesian method updates a posterior over the perceptual bias after each trial. The prior can be specified based on data from past sessions, thus reducing the variability of the online estimates and allowing it to converge to a stable estimate over a small number of trials. After validating our method on synthetic data, we apply it to estimate perceptual biases of monkeys in a motion direction discrimination task in which varying background optic flow induces robust perceptual biases. This method overcomes an important challenge to understanding the neural basis of subjective percepts.

HiTSKT: A hierarchical transformer model for session-aware knowledge tracing

Knowledge tracing (KT) aims to leverage students’ learning histories to estimate their mastery levels on a set of pre-defined skills, based on which the corresponding future performance can be accurately predicted. In practice, a student’s learning history comprises answers to sets of massed questions, each known as a session, rather than merely being a sequence of independent answers. Theoretically, within and across these sessions, students’ learning dynamics can be very different. Therefore, how to effectively model the dynamics of students’ knowledge states within and across the sessions is crucial for handling the KT problem. Most existing KT models treat student’s learning records as a single continuing sequence, without capturing the sessional shift of students’ knowledge state. To address the above issue, we propose a novel hierarchical transformer model, named HiTSKT, comprises an interaction(-level) encoder to capture the knowledge a student acquires within a session, and a session(-level) encoder to summarize acquired knowledge across the past sessions. To predict an interaction in the current session, a knowledge retriever integrates the summarized past-session knowledge with the previous interactions’ information into proper knowledge representations. These representations are then used to compute the student’s current knowledge state. Additionally, to model the student’s long-term forgetting behaviour across the sessions, a power-law-decay attention mechanism is designed and deployed in the session encoder, allowing it to emphasize more on the recent sessions. Extensive experiments on four public datasets demonstrate that HiTSKT achieves new state-of-the-art performance on all the datasets compared with seven state-of-the-art KT models.

Multi-Modal Domain Adaptation Variational Autoencoder for EEG-Based Emotion Recognition

Traditional electroencephalograph (EEG)-based emotion recognition requires a large number of calibration samples to build a model for a specific subject, which restricts the application of the affective brain computer interface (BCI) in practice. We attempt to use the multi-modal data from the past session to realize emotion recognition in the case of a small amount of calibration samples. To solve this problem, we propose a multi-modal domain adaptive variational autoencoder (MMDA-VAE) method, which learns shared cross-domain latent representations of the multi-modal data. Our method builds a multi-modal variational autoencoder (MVAE) to project the data of multiple modalities into a common space. Through adversarial learning and cycle-consistency regularization, our method can reduce the distribution difference of each domain on the shared latent representation layer and realize the transfer of knowledge. Extensive experiments are conducted on two public datasets, SEED and SEED-IV, and the results show the superiority of our proposed method. Our work can effectively improve the performance of emotion recognition with a small amount of labelled multi-modal data.

Can integrating the Memory Support Intervention into cognitive therapy improve depression outcome? A randomized controlled trial

We investigated if improving a patient's memory for the content of their treatment, via the Memory Support Intervention, improves illness course and functional outcomes. The platform for investigating this question was major depressive disorder (MDD) and cognitive therapy (CT). Adults diagnosed with MDD (N = 178) were randomly allocated to CT + Memory Support (n = 91) or CT-as-usual (n = 87). Both treatments were comprised of 20–26, 50-min sessions over 16 weeks. Blind assessments were conducted before and immediately following treatment (post-treatment) and 6 months later (6FU). Patient memory for treatment, assessed with a free recall task, was higher in CT + Memory Support for past session recall at post-treatment. Both treatment arms were associated with reductions in depressive symptoms and functional impairment except: CT + Memory Support exhibited lower depression severity at 6FU (b = -3.09, p = 0.050, d = −0.27), and greater reduction in unhealthy days from baseline to 6FU (b = −4.21, p = 0.010, d = −1.07), compared to CT-as-usual. While differences in illness course and functional outcomes between the two treatment arms were limited, it is possible that future analyses of the type of memory supports and longer follow-up may yield more encouraging outcomes. Trial registrationClinicalTrials.gov NCT01790919. Registered October 6, 2016.

TR-GAN: Multi-Session Future MRI Prediction With Temporal Recurrent Generative Adversarial Network

Magnetic Resonance Imaging (MRI) has been proven to be an efficient way to diagnose Alzheimer's disease (AD). Recent dramatic progress on deep learning greatly promotes the MRI analysis based on data-driven CNN methods using a large-scale longitudinal MRI dataset. However, most of the existing MRI datasets are fragmented due to unexpected quits of volunteers. To tackle this problem, we propose a novel Temporal Recurrent Generative Adversarial Network (TR-GAN) to complete missing sessions of MRI datasets. Unlike existing GAN-based methods, which either fail to generate future sessions or only generate fixed-length sessions, TR-GAN takes all past sessions to recurrently and smoothly generate future ones with variant length. Specifically, TR-GAN adopts recurrent connection to deal with variant input sequence length and flexibly generate future variant sessions. Besides, we also design a multiple scale & location (MSL) module and a SWAP module to encourage the model to better focus on detailed information, which helps to generate high-quality MRI data. Compared with other popular GAN architectures, TR-GAN achieved the best performance in all evaluation metrics of two datasets. After expanding the Whole MRI dataset, the balanced accuracy of AD vs. cognitively normal (CN) vs. mild cognitive impairment (MCI) and stable MCI vs. progressive MCI classification can be increased by 3.61% and 4.00%, respectively.

충남대학교 박물관 소장 전곡리 석기들의 고찰

This article is a substitute for an official report of the lithic assemblage stored at Chungnam National University Museum (CNUM). The assemblage was originally acquired by previous museum personnel through several field trips at Jeongokri in 1979. Taking advantage of introducing the Jeongokri assemblage at CNUM, we also reviewed and examined the result of the past Jeongokri Excavation sessions during the earliest phase from 1979 to 1981 in order to suggest some future issues about the archaeological research of the Jeongokri site. A total of 196 lithic specimens were observed and re-classified according to a more refined typological scheme; their quantitative attributes were compiled and analyzed with the same method as adapted on the previously excavated Jeongok ACF assemblage. The result is that: 1) even though the CNUM Jeongokri assemblage is of very cursory and fragmentary surface collection, the assemblage characteristics are quite similar to those of other previously excavated Jeongokri assemblages, 2) and that more clearly defined lithic categories can be established and a new tool type—the pestle—can be suggested to successfully replace the previous ambiguous type—the handplane. In addition, we closely evaluated the context of discovered handaxes within sediments and approached the fundamental reason why the age of the Jeongokri handaxe was wrongfully understood and miscalculated so far. Taking an example of the recently discovered new Jeongokri 85-12 locality, we summoned a caution that the age estimation of the Jeongokri assemblage need to be clearly based on more rigorous designation of typologically genuine handaxes, as well as on the sound explanation of artifact horizon and relevant formation processes lest any more previous trial-and-error should emerge again.

Codevelopment Action Learning during the pandemic – findings from two online co-learning and co-creation events

ABSTRACT This article discusses the international events known as Proxima, organized by a group of colleagues, whose slogan is ‘Drawing Closer to Go Further’. Two half-day sessions, held in April and May 2020, consisted of 20 simultaneous Codevelopment Group (CDG) online sessions. Held during the global lockdown, these online events generated creative, collaborative and transformative spaces for participants throughout the French-speaking world. A total of 148 people from nine countries, including 15 experienced facilitators took part in this unique experience. After giving an overview of the Action Learning-based CDG method, this article presents the survey results on participant goals, takeaways and the intent to apply/transfer, and shows that 96% of participants responded positively to questions about these aspects. Qualitative content analyses also shows what was achieved, learned and could be transferred, such as a better understanding of the CDG method and its online applications. Overall, this experience further reinforces the fact that CDGs can effectively stimulate meaningful learning, creativity and collaborative action. Finally, in the past, Action Learning and CDG sessions have tended to be held in-person. However, these events showed that virtual sessions can be effective and generate impact when physical presence is not possible.

Delegation Using Forward Induction

This paper explores how delegation can be used as a signal to sustain cooperation. I consider a static principal–agent model with two tasks, one resembling a coordination game. If there is asymmetric information about the agent’s type, the principal with high private belief can delegate the first task as a signal. This is also supported by the forward induction argument. However, in the laboratory setting, this equilibrium is chosen only sometimes. When the subjects have information about past sessions, there is a significant increase in the use of delegation. This finding sheds light on equilibrium selection in Bayesian games.

Impact of Digital and Non-Digital Urban Participatory Approaches on Public Access Conditions: An Evaluation Framework

The gradual institutionalization of public participation increasingly compels local authorities to partially share their power over the transformation of urban areas. The smooth running of a participatory session is based on selecting the appropriate type of interaction, or medium, which supports the local authorities to reach and interact with a targeted public. However, local authorities often appear unfamiliar with the organization of interactive sessions with the population. This article introduces an evaluation framework that focuses on the access conditions of participants to the sessions of interaction. This novel perspective aspires to assist the local authorities in their decision to adopt a participatory medium (or method of interaction). Seven dimensions are investigated to this aim, namely accessibility, availability, adequacy, affordability, acceptability, awareness, and attractiveness (the last dimension is introduced in this article). In light of two real case scenarios that occurred in Western Switzerland, the use of the access framework is investigated for two potential purposes: (1) supporting the choice of a medium for an interactive session according to the urban project’s context and the targeted public; and (2) improving future participatory approaches by assessing the representativeness of participants attending a past session in comparison to the originally targeted public.

More than just the mental health act – foundation-specific teaching to inspire psychiatrists

AimsThis project aimed to further develop a teaching programme for Foundation Doctors attached to a psychiatry rotation. The purpose was threefold – to educate foundation doctors about important psychiatric topics; to encourage them to think about wider impacts of psychiatry; and to inspire them to consider psychiatric training in the longer term.BackgroundThe Royal College of Psychiatrists’ mission statement includes actively promoting psychiatry as a career and improving knowledge of mental health, including its interactions with people's physical and social backgrounds. Targeting foundation doctors rotating into psychiatry posts is a good opportunity to achieve these objectives, as they will be the cross-speciality doctors of the future, and have specific learning needs given their unique rotations and new medical careers.MethodOn one Wednesday morning per month Foundation Doctors had a specific teaching session for them. The sessions consisted of four 30-minute teaching blocks which, crucially, were given by foundation doctors. They were facilitated by a core psychiatry trainee, and the topics were decided by the doctor teaching each 30-minute block. The foundation doctors were able teach on any topic related to psychiatry that interested them. Feedback forms were developed and provided at the end of each session for the foundation doctors, as well as at the end of each recent foundation rotation, to get feedback on the overall quality of the course delivered.ResultThe programme has now had 6 complete cohorts of foundation doctors. We have built a varied topic bank from past sessions, including the Mental Health Act, dementia, the Mind-Body Problem, psychiatry in video games and sociology of psychiatric illness, amongst other topics. All foundation doctors questioned have agreed or strongly agreed that the sessions were helpful for their psychiatric rotation and general medical training. Particularly praised aspects were the ability to discuss psychiatric topics that weren't normally discussed in an academic environment, being able to take ownership over learning and practicing giving teaching. Vitally, core trainee facilitators also found the sessions inspiring for their training.ConclusionThe Foundation Teaching Programme has increased doctors’ knowledge of a range of psychiatric topics, the breadth of which and agency in choosing topics has increased engagement with psychiatry, regardless of planned medical training speciality. Areas to explore in the future include potentially opening attendance to medical students and physician associate students, and to other regions of the deanery. Evaluating the long-term impact of this training is also warranted.

I-CARE-An Interaction System for the Individual Activation of People with Dementia.

I-CARE is a hand-held activation system that allows professional and informal caregivers to cognitively and socially activate people with dementia in joint activation sessions without special training or expertise. I-CARE consists of an easy-to-use tablet application that presents activation content and a server-based backend system that securely manages the contents and events of activation sessions. It tracks various sources of explicit and implicit feedback from user interactions and different sensors to estimate which content is successful in activating individual users. Over the course of use, I-CARE’s recommendation system learns about the individual needs and resources of its users and automatically personalizes the activation content. In addition, information about past sessions can be retrieved such that activations seamlessly build on previous sessions while eligible stakeholders are informed about the current state of care and daily form of their protegees. In addition, caregivers can connect with supervisors and professionals through the I-CARE remote calling feature, to get activation sessions tracked in real time via audio and video support. In this way, I-CARE provides technical support for a decentralized and spontaneous formation of ad hoc activation groups and fosters tight engagement of the social network and caring community. By these means, I-CARE promotes new care infrastructures in the community and the neighborhood as well as relieves professional and informal caregivers.

Service Learning Opportunity for Undergraduate Science Students: Integrating Problem-Based Learning in the High School Science Curriculum

A service learning course has been developed through which undergraduate science students create problem-based learning materials for use in high school science classes. Undergraduate students research active learning pedagogies with a specific focus on problem-based learning, create original materials to be used in grades 11 and 12 chemistry and environmental science classes, and work in coordination with secondary school teachers to facilitate the use of their original problems in high school classes. For many of the participating students, this is their first experience with self-directed learning and a rare opportunity to develop pedagogical content knowledge. Beginning through a partnership with a single high school teacher, this program has expanded to include nine teachers at five different schools, and interest continues to grow. This article presents the motivations and goals of the course for participating students, the structure and outline of the course, sample works that have been created by students, as well as some student outcomes from past sessions of the course.

An Efficient Blacklistable Anonymous Credentials without TTP of Tracing Authority Using Pairing-Based Accumulator

In conventional ID-based user authentications, privacy issues may occur, since users’ behavior histories are collected in Service Providers (SPs). Although anonymous authentications such as group signatures have been proposed, these schemes rely on a Trusted Third Party (TTP) capable of tracing misbehaving users. Thus, the privacy is not high, because the TTP of tracing authority can always trace users. Therefore, the anonymous credential system using a blacklist without the TTP of tracing authority has been proposed, where blacklisted anonymous users can be blocked. Recently, an RSA-based blacklistable anonymous credential system with efficiency improvement has been proposed. However, this system still has an efficiency problem: The data size in the authentication is O(K0), where K0 is the maximum number of sessions in which the user can conduct. Furthermore, the O(K0)-size data causes the user the computational cost of O(K0) exponentiations. In this paper, a blacklistable anonymous credential system using a pairing-based accumulator is proposed. In the proposed system, the data size in the authentication is constant for parameters. Although the user’s computational cost depends on parameters, the dependent cost is O(δBL · K) multiplications, instead of exponentiations, where δBL is the number of sessions added to the blacklist after the last authentication of the user, and K is the number of past sessions of the user. The demerit of the proposed system is O(n)-size public key, where n corresponds to the total number of all sessions of all users in the system. But, the user only has to download the public key once.

Estimating and predicting athlete performance using Machine Learning

A great amount of aspects of cycling could benefit from the revolutionary solutions available nowadays thanks to data science. This advanced computational discipline could be used to model sorely complex physiological processes, such as that of the human body, thus might have a key role in physical performance. In this article, we introduce how Machine Learning algorithms are used to estimate the current physical abilities of an athlete (i.e. his maximal power output) or predict his performance along the next few weeks. As pointed out onto Figure 1, the various features required for our Machine learning model are determined from already existing labelled data during a phase called training , based on an archive containing any cycling activity from a given rider. For each session, we consider several computed average metrics (power, heart rate, cadence…), as well as the ones quantifying the session workload and the level of shape of the athlete. Once the model is said to be trained, it eventually makes it possible to estimate the rider’s current power profile. Regarding prediction , the model is first trained over a dataset containing an archive of past sessions, compiled with the upcoming ones, planned by the athlete training program. This allows us to predict the peak power output of the athlete over the next few weeks. In order to validate our approach, we used our algorithms to estimate and predict the maximal 10-minute peak power of a world-class athlete over 3 months. The results we obtained through our models are displayed on Figure 2. These results should be compared with the all-out test values (red dots) produced by the rider every week. The solid lines show power values estimations generated by two different algorithms, calculating the current shape of the rider (i.e. with models trained over past sessions only). The two models are based on different estimations of the workload and the level of shape. The basic estimation model (fuchsia line) used the Training Stress Score and the Training Stress Balance while the advanced estimation model (light blue line) is based on metrics normalized to the athlete. Therefore, those metrics vary between 0% and 100%. As can be observed on Figure 2, the advanced model is significantly more accurate than the basic one in approaching the values extracted from the all-out tests, standard deviations obtained with these models being quantified respectively at 8W and 22 W. Regarding predictions (blue dashed line), we take the sessions originally planned by the rider’s training program into account in order to train our model. The figures obtained thanks to the advanced predicting model still match the values from the all-out tests pretty well, as pointed out by a tiny 10W standard deviation. As shown above, the results obtained in this article highlight the accuracy and reliability achievable thanks to advanced data science, which proves to be the current best tool to estimate any current peak power output of a given athlete, and to calculate his performance during the near future. This approach sets the path of a great amount of applications, especially for determining and optimizing a custom training plan, allowing any athlete to reach his maximal level of performance at the very day he/she chose to.

P057: Evidence-based medicine (EBM) simulation: teaching real-time literature searching to emergency medicine residents using a flipped classroom and high-fidelity simulation

Innovation Conept: Evidence-based medicine (EBM), including literature search skills, is an objective of the Emergency Medicine (EM) residency curriculum. Traditional teaching of this topic utilized a classroom-based, librarian-lead session that presented an overview of many search engines. Feedback from past sessions indicated that learners retained little after the session. To be effective, EBM needs to be brought to the bedside. We created a session to engage EM residents and improve their efficiency in literature searching during an EM shift. Methods: We conducted a needs assessment among EM residents in our program. In response to this and to maximize impact of teaching, we created an EBM workshop on literature searching that used a flipped classroom approach and high-fidelity simulation. The session was designed for a small group (12 junior residents), with the goals of being interactive, engaging and practice-relevant. Feedback was collected on the simulation experience. Curriculum, Tool or Material: With a librarian, we created a brief list of EM-relevant databases. It included tips for searching and links to the corresponding sites / apps. Students received the list 7 days prior and were instructed to set up the resources on their smartphones. Pre-readings also covered the hierarchy of evidence and formulating a good clinical (PICO) question. All students participated in the high-fidelity simulation, with one volunteer leader. The case involved a stable patient. Residents proceeded with initial case assessment until they faced a management decision that required a literature search. All residents participated on their smart phones. Collectively, it took 5 minutes to find a study that adequately addressed the clinical question. The patient was managed accordingly and symptoms resolved. Feedback on the simulation was abundantly positive. Students found it engaging, practical and realistic. It helped them learn to efficiently search the literature while managing a stable patient. Conclusion: Using a multi-modal teaching strategy that includes simulation makes teaching EBM literature searching more interesting, engaging and applicable to EM practice. Future work will look at creating further sessions to reinforce and promote retention of key concepts and integrate them into EM practice.

Mutual healing enabled group-key distribution protocol in Wireless Sensor Networks

The unreliable wireless medium for communication within a Wireless Sensor Network (WSN) necessitates equipping the sensor nodes with shared secrets to protect the communication en-route. In order to establish a dynamic session secret amongst a group of sensor nodes in a WSN, group-key broadcast is found to be a workable solution. However, when a node misses one or more group-key broadcast messages, rebroadcasting results in additional energy consumption at node end and increased network traffic as well. Self-healing mechanism is, therefore, used to recover the lost broadcasts using a future session broadcast. A node, that has missed multiple consecutive broadcasts and does not want to wait for a future broadcast, can obtain the session key through mutual-healing with the help of a neighbor node. In this paper, we present a secure and efficient mutual-healing protocol that uses Chinese remainder theorem(CRT) based secret sharing for group key broadcast. During mutual-healing, mutual authentication and key confirmation are provided. The probability of an unauthorized member retrieving a past session key through self-healing or responding to a mutual-healing request as a neighbor node is shown negligible. The experimental results show that the proposed protocol provides significant performance improvement in terms of computation, communication and storage overhead as compared to existing mutual-healing protocols.

Collaborative Filtering using Euclidean Distance in Recommendation Engine

Objectives: Recommendation engine is a part of information filtering system that tries to predict the ‘preference’ or ‘rating’ of an item in the E-commerce. Recommendation engines have become extremely common in recent days to make an appropriate recommendation rapidly and effectively about any products on which the user is interested. Methods/ Statistical Analysis: One of popular information filtering systems in the recommendation engine is collaborative filtering where the predictions are made based on the usage patterns of the users who are similar to another user. The accuracy of a recommendation engine using collaborative filtering depends on the techniques used to measure the similarity between the user’s preferences. Therefore, in this paper we use two metrics to measure the similarity between the user’s preferences namely KL Divergences and Euclidean distance. The proposed algorithm works by first clustering the users using k means clustering by utilising the similarity metrics and then computing the global Markov matrix for that cluster. Next, the PageRank value for each user is computed and those values are combined with the global Markov matrix to find the recommendations. Findings: We consider the problem of collaborative filtering to recommend potential items of interest to a user already engaged in a session, using past session of the user and other users. Our algorithm leads to the personalized PageRank, where context is captured by the personalization vector. The results show that the collaborative filtering using Euclidean distance metrics for similarity measure performs well than the KL divergence. Application/ Improvements: The proposed recommendation engine can be used in a wide variety of applications such music, movies, books, news, research articles, social media, search queries, and products in general in order to provide a effective recommendation.

A Provably Password Authenticated Key Exchange Scheme Based on Chaotic Maps in Different Realm

Until now, the overwhelming majority of password- authenticated key agreement protocols using chaotic maps are based on three architectures (client/server, two clients/server and multi-server) and four security models (heuristic security, random oracle, ideal cipher and standard model). However, with rapid changes in the modern communication environment such as wireless mesh net- works and cloud storing, it is necessary to put forward a kind more flexible and general architecture to adapt it. So, in our paper, we firstly propose a provable secure two-party in two-realm key agreement protocol using chaotic maps in the standard model. Our proposed protocol is more general and it is easy to expand to many other forms, such as three-party or N-party in different realms. The new protocol resists dictionary attacks mounted by either passive or active network intruders, allowing, in principle, even weak password phrases to be used safely. It also offers perfect forward secrecy, which protects past sessions and passwords against future compromises. Finally, we give the security proof in the standard model and the efficiency analysis of our proposed scheme.

Classification-based prediction models for stock price index movement

Stock price prediction with high accuracy may offer significant opportunities for the investors who make decisions on making profit or having high gains over the stocks in stock markets. In this study, four predictive models have been developed for classification task in predicting the direction of movement in the sessional, daily, weekly, and monthly Istanbul Stock Exchange National (ISEN) 100 Index using five years of data. Multilayer perceptron (MLP), which comprises artificial neural networks (ANN), Logistic Regression (LR), and Bagging of Logistic Regression (BLR) classification techniques are used in the models. During the prediction, four datasets are used and the following factors are taken into account: data of macroeconomic indicators, gold prices, oil prices, foreign exchange prices, stock price indexes in various countries, and the data of the ISEN 100 index for past sessions and prior days, which are used as input variables in the datasets. In connection with that, the most effective factors of these input variables were determined by using some feature selection methods. As a result, prediction performances showed that using reduced datasets consisting of only selected the most important features induced a predictive model of each dataset for classification modelling with a better predictive accuracy than using original datasets. Experimental results showed that prediction performances of the models, which are 64.13%, 63.09%, 81.54%, and 100% for the sessional, daily, weekly, and monthly datasets respectively, were found by MLP significantly better than the other classifiers used in this study.