Kean Digital Learning Commons
Not a member yet
8043 research outputs found
Sort by
A Hybrid Model Based on Multi-LSTM and ARIMA for Time Series Forcasting
No full textIn recent years, deep learning has rapidly developed and been widely applied across different fields. While statistical models are known for their powerful interpretability and unique processing mechanisms, time series are a common form of data that have found widespread use in areas such as economics, finance, and environmental science. The primary goal of time series analysis is to reveal the laws and characteristics of the data to make accurate predictions and informed decisions. However, since different methods have their own advantages and disadvantages, hybrid models can effectively combine them to expand the benefits and reduce the drawbacks. This paper proposes a time series forecasting model based on a multivariate LSTM with ARIMA. The ARIMA method is first used to analyze the time series data and obtain prediction results and confidence intervals. Then, these results are combined with other multivariate variables to serve as input features for the multivariate LSTM model to predict and model the time series data. The study selects the opening, closing, high, and low prices of Apple stock in the United States from March 1, 2022, to March 1, 2023, for prediction data. And the prices from March 1, 2023 to March 15, 2023 are for the following test. The results demonstrate that the proposed method is more effective than traditional methods
Exploring drought tolerance in melon germplasm through physiochemical and photosynthetic traits
No full textDrought stress is a global concern that has a negative impact on the growth and production of melon (Cucumis melo L.). In this study, 58 melon accessions were subjected to drought stress induced by polyethylene glycol (PEG-6000). Comprehensive evaluations were performed to identify various morphological, biochemical, and physiological attributes of melon. Drought stress significantly reduced shoot length (SL), stem diameter (SD), leaf width (LW), and leaf length (LL) in the melon seedlings. Similarly, drought stress resulted in a significant reduction in photosynthetic pigments (Chl, Car), relative water content (RWC), chlorophyll fluorescence (Fv/Fm), transpiration rate (Tr), stomatal conductance (Gs), net photosynthetic rate (Pn) and intercellular CO2 concentration (Ci). On the other hand, biochemical indicators such as malondialdehyde content (MDA), soluble protein content (SP) and soluble sugar content (SS) were observed to be enhanced upon exposure to drought stress. Most indicators showed strong positive correlations based on Pearson correlation analysis. Furthermore, the modified membership function and D values for the drought tolerance indices were calculated to evaluate the drought tolerance level of melon accessions. In addition, melon accessions were classified into drought-resistant and drought-sensitive groups based on cluster analysis. As a result, mel-46, mel-58 and mel-15 were identified as drought-resistant genotypes among the assessed melon accessions. Taken together, these accessions provide potential genetic resources for further improvement and breeding of melon genotypes. Furthermore, the indicators responsible for the assessment of drought tolerance can provide a baseline for future studies
Model comparison of regression, neural networks, and XGBoost as applied to the English Premier League transfer market
No full textThe English Premier League (EPL) is the highest level in the UK professional soccer system and one of the largest and most competitive professional soccer leagues in the world. This research examines factors influencing transfer fees in the most popular transfer market for EPL using data spanning ten years. By building a Nash equilibrium model, a dynamic modelling system is developed to measure the transfer fee with the ordinal least square regression, eXtreme Gradient Boosting (XGBoost), and neural network (NN) models. The study recognises the effect of bargaining power and provides optimised strategies for clubs and players. Moreover, clubs will be able to utilise NN as the most advantageous method to determine the transfer fees. The research can serve as a comprehensive decision support system for assessing the expenditure needs corresponding to the players scouted in the transfer market
PointNorm: Dual Normalization is All You Need for Point Cloud Analysis
No full textPoint cloud analysis is challenging due to the irregularity of the point cloud data structure. Existing works typically employ the ad-hoc sampling-grouping operation of PointNet++, followed by sophisticated local and/or global feature extractors for leveraging the 3D geometry of the point cloud. Unfortunately, the sampling-grouping operations do not address the point cloud\u27s irregularity, whereas the intricate local and/or global feature extractors led to poor computational efficiency. In this paper, we introduce a novel DualNorm module after the sampling-grouping operation to effectively and efficiently address the irregularity issue. The DualNorm module consists of Point Normalization, which normalizes the grouped points to the sampled points, and Reverse Point Normalization, which normalizes the sampled points to the grouped points. The proposed framework, PointNorm, utilizes local mean and global standard deviation to benefit from both local and global features while maintaining a faithful inference speed. Experiments show that we achieved excellent accuracy and efficiency on Model-Net40 classification, ScanObjectNN classification, ShapeNetPart Part Segmentation, and S3DIS Semantic Segmentation. Code is available at https://github.com/ShenZheng2000/PointNorm-for-Point-Cloud-Analysis
An Artificial Intelligence-Based Interactive Learning Platform to Assist Visually Impaired Children in Learning Mathematics
No full textVisually impaired children mainly depend upon hearing and touch in the absence of vision. Smartphones are now relatively cheap and are in widespread use in almost all parts of the world, including by many people who do not have access to laptops or desktop machines. Smartphone-based applications provide a learning environment in which visually impaired children can enhance their educational skills in a similar way to other students. This paper introduces an artificial intelligence-based interactive learning platform that can enhance the mathematical skills of visually impaired children. This platform can assist teachers in the classroom to provide accessible and interactive materials to their visually impaired students. The proposed platform uses text-to-speech along with vibrotactile and auditory feedback to help visually impaired students arrive at a better understanding of mathematical material. Four participants were recruited to evaluate the prototype of this interactive learning platform. The results showed that understanding of mathematical content in visually impaired children was significantly improved. Furthermore, problem-solving skills and action awareness were enhanced with the help of a multimodal feedback approach. Participants also reported high levels of satisfaction with the proposed design. The paper concludes with a discussion of possible directions for future research aimed at overcoming barriers to learning faced by visually impaired children
Attention-Based Model for Sentiment Analysis
No full textWith the evolution of deep learning, work has been carried out in various fields of NLP. With sentiment analysis, the reviews generated on various platforms can be easily analyzed and could be classified based on the polarity. In this work, IMDB movie dataset has been used and results are compared with the existing baseline techniques. A model is implemented using attention-based mechanism. BiLSTM is used in extracting the global features from the text and for overcoming the gradient issues. This model helps in understanding the contextual relationship in the words just like the human brain. With the use of attention model, the problems of long-term dependencies are resolved thus giving better performance. Evaluation parameters are computed based on the accuracy, precision, recall, and F-score. Results of this model prove to be better than the previous existing techniques
Can renewable energy microfinance promote financial inclusion and empower the vulnerable?
No full textEnergy microfinance is a promising asset for ensuring access to finance and electricity for the poor and women living in rural areas. Targeting vulnerable categories, renewable energy microfinance tools aim to promote sustainability and resilience policies to foster sustainable energy and ecological transitions. This research note explores the role of energy microfinance policies in facilitating virtuous loops for empowering the vulnerable, enhancing entrepreneurship, fighting poverty, and promoting social inclusion and environmental protection – as prescribed by the sustainable development mandate. Examining the microfinance-renewable energy-vulnerable nexus, the paper proposes a conceptual contribution and fungible development policy recommendations, sketching the main drivers of this linkage. To this end, a critical review of existing publications and successful world experiences is proposed. The analysis finds out a relevant potential for energy microfinance, whereby renewable energy expansion can be a strong player in shaping fresh microfinance strategies and prospects, providing ecological, social, economic and governance benefits and new research, policy, and practical agenda
Turkish coffee bean imports: asymmetric exchange rate pass-through analysis
No full textPurpose: As one of the world\u27s most valuable traded commodities, the market for coffee beans has grown enormously in recent years. The paper aims on analyzing the nonlinear exchange rate pass-through in Turkish coffee bean imports from two important sources in South America: Brazil and Colombia. Design/methodology/approach: Data collected in this paper through reliable channels include nominal import value, exchange rate, production of total industry, etc. Independent and dependent variables are obtained through conversion. Since the nonlinearly adjusted exchange rate differs significantly from the linearly adjusted one for the export trade of Brazilian coffee beans, this paper develops the autoregressive distributed lag (ARDL) and nonlinear ARDL frameworks and demonstrates their application through asymmetric cointegration and error correction models. Findings: The results of this paper show that imports of Brazilian coffee bean exhibit a more dramatic asymmetry compared to Colombia\u27s coffee bean imports. The results of this study contribute to the import trade of non-oil commodities in developing countries, particularly Brazil, and enrich the existing literature on nonlinear exchange rate adjustments. Research limitations/implications: The export of Colombian coffee beans is not as old as Brazil, and it was not until much later that Colombia began to export coffee beans to the rest of the world. Originality/value: The present study is an addition to the literature of agricultural trade. The authors analyze the nonlinear exchange rate pass-through in Turkish coffee bean imports from two important sources in South America: Brazil and Colombia. Different from the current mainstream research on oil commodity trade, this paper focuses on international trade from the perspective of coffee beans, which can enlighten the practice in this field