Oliver’s four-factor model: Validation through causality

While there has been a growing interest in sports analysis in recent years, much research first focused on a classical statistical approach and later on an artificial intelligence approach. This article aims instead to propose a causal inference approach to sports analysis. In particular, the present article intends to review the famous four-factor model proposed by Dean Oliver for assessing the winning ability of National Basketball Association (NBA) teams through a causal inference approach. A structural equation model is used to validate Oliver’s model. The present paper considers the winning percentage and the factors’ statistics over entire seasons from [Formula: see text] to [Formula: see text]. The statistics for the [Formula: see text] season are considered only on a subset of the games. This is because the games played in the Orlando bubble under the particular COVID-19 situation have been regarded as outliers compared to the games played in the other NBA seasons, hence they have not been taken into account. The second goal of the article is to analyse if the fitting ability of the four-factor model changes when it is fitted over the pre[Formula: see text] and post[Formula: see text] basketball eras datasets, considering the year [Formula: see text] as the turning point for the NBA playing style.

VOVU: A Method for Predicting Generalization in Deep Neural Networks

During the development of deep neural networks (DNNs), it is difficult to trade off the performance of fitting ability and generalization ability in training set and unknown data (such as test set). The current solution is to reduce the complexity of the objective function, using regularization methods. In this paper, we propose a method called VOVU (Variance Of Variance of Units in the last hidden layer) to maximize the optimization of the balance between fitting power and generalization during monitoring the training process. The main idea is to give full play to the predictability of the variance of the hidden layer units in the complexity of the neural network model and use it as a generalization evaluation index. In particular, we take full advantage of the last layer of hidden layers since it has the greatest impact. The algorithm was tested on Fashion-MNIST and CIFAR-10. The experimental results demonstrate that VOVU and test loss are highly positively correlated. This implies that a smaller VOVU indicates that the network has better generalization. VOVU can serve as an alternative method for early stopping and a good predictor of the generalization performance in DNNs. Specially, when the sample size is limited, VOVU will be a better choice because it does not require dividing training data as validation set.

Prediction And Evaluation of Forest Fire In Yunnan of China Based On Geographically Weighted Logistic Regression Model

Establishing an effective forest fire forecasting mechanism is the premise of scientific planning and management of forest fires and forest fire prevention. In recent years, the forest fire prediction mechanism has been one of the key areas of concern for the government forestry management departments and forestry researchers. One of them, is logistic regression ( LR ). It is a relatively frequent prediction probability model used in forest fire prediction and prediction in China and abroad for the past few years. However, with the gradual deepening of research, it is found that the logistic regression model fails to fully consider the spatial non-stationary relationship between forest fires and driving factors, which leads to poor fitting effect and low prediction accuracy of the model. But its extended counterpart, the Geographically weighted logistic regression ( GWLR ) model, takes into account the spatial correlation between model variables, and effectively improves the fitting ability and prediction accuracy of the model. Therefore, this paper compares the ability of the logistic regression model and the geographically weighted logistic regression model in terms of fitting ability and prediction accuracy in order to obtain the ability of the two models to predict forest fires in Yunnan Province. In this paper, the samples were divided into 60% training samples and 40% test samples, and repeated sampling was carried out 5 times for training. Variables that appeared in the training model for 3 or more times were used to construct the final LR and GWLR models. Finally, the models with better fitting ability and higher prediction accuracy were used to classify the fire risks in Yunnan Province. The results show that the geographically weighted logistic regression model is superior to the logistic regression model in terms of fitting effect and accuracy. The geographically weighted logistic regression model is more suitable for the data structure of forest fires in Yunnan Province and has better prediction ability. The AUC value of the geographically weighted logistic regression model is 0.902, and the prediction accuracy is 82.7 %; The AUC value of logistic regression model was 0.891, and the prediction accuracy was 80.1%; Fully considering the spatial heterogeneity among model variables can, to some extent, predict forest fires more accurately. The fitting of the two models shows that the relative humidity, temperature, air pressure, sunshine hours, daily precipitation, wind speed, and other meteorological factors; Vegetation type; terrain factor; Population density, road network and other human activity factors become the cause of forest fires in Yunnan Province.

An Intelligent Quadrotor Fault Diagnosis Method Based on Novel Deep Residual Shrinkage Network

In this paper, a fault diagnosis algorithm named improved one-dimensional deep residual shrinkage network with a wide convolutional layer (1D-WIDRSN) is proposed for quadrotor propellers with minor damage, which can effectively identify the fault classes of quadrotor under interference information, and without additional denoising procedures. In a word, that fault diagnosis algorithm can locate and diagnose the early minor faults of the quadrotor based on the flight data, so that the quadrotor can be repaired before serious faults occur, so as to prolong the service life of quadrotor. First, the sliding window method is used to expand the number of samples. Then, a novel progressive semi-soft threshold is proposed to replace the soft threshold in the deep residual shrinkage network (DRSN), so the noise of signal features can be eliminated more effectively. Finally, based on the deep residual shrinkage network, the wide convolution layer and DroupBlock method are introduced to further enhance the anti-noise and over-fitting ability of the model, thus the model can effectively extract fault features and classify faults. Experimental results show that 1D-WIDRSN applied to the minimal fault diagnosis model of quadrotor propellers can accurately identify the fault category in the interference information, and the diagnosis accuracy is over 98%.

Return Prediction Based on Discriminating market-styles with Reinforcement Learning

For hedge funds, return prediction has always been a fundamental and important problem. Usually, a good return prediction model directly determines the performance of a quantitative investment strategy. However, the performance of the model will be influenced by the market-style. Even the models trained through the same data set, their performance is different in different market-styles. Traditional methods hope to train a universal linear or nonlinear model on the data set to cope with different market-styles. However, the linear model has limited fitting ability and is insufficient to deal with hundreds of features in the hedge fund features pool. The nonlinear model has a risk to be over-fitting. Simultaneously, changes in market-style will make certain features valid or invalid, and a traditional linear or nonlinear model is not sufficient to deal with this situation. This thesis proposes a method based on Reinforcement Learning that automatically discriminates market-styles and automatically selects the model that best fits the current market-style from sub-models pre-trained with different categories of features to predict the return of stocks. Compared with the traditional method that training return prediction model directly through the full data sets, the experiment shows that the proposed method has a better performance, which has a higher Sharpe ratio and annualized return.

Bearing Fault Identification Based on Deep Convolution Residual Network

Bearings are important parts in industrial production and are related to the normal operation of mechanical equipment. For bearing fault identification， traditional method often includes feature extraction, which involves professional prior knowledge and is time-consuming. This paper proposes the deep convolution residual network (DCRN) to identify the bearing fault. Based on the end-to-end learning characteristics of deep neural networks, this method can directly use raw data for training, and does not require feature extraction. Moreover, under the effect of skip connection, DCRN can exert the powerful fitting ability of deep neural network. In this paper, by stacking residual blocks, three different architecture of DCRN are designed and all three achieve very high accuracy, respectively 99.60%, 99.71% and 99.81%. Compared with other methods, DCRN have better generalization performance.

Screening and Validation of a Novel T Stage-Lymph Node Ratio Classification for Colon Cancer

Purpose Lymph node ratio (LNR) has advantages in predicting prognosis over the American Joint Committee on Cancer (AJCC) N stage. However, the prognostic value of establishing a novel T stage-Lymph Node Ratio classification (TLNR) for colon cancer by combining LNR and T stage is currently unknown.Methods We included 62,294 stage I-III colon cancer patients from the SEER data base as a training cohort. An external validation was performed in 3,327 additional patients. A novel LNR stage was established and included into a novel TLNR classification by combining with T stage. Patients with similar survivals were grouped according to T and LNR stages, with T1LNR1 as a reference. Results We developed a novel TLNR classification: stages I (T1LNR1-2, T1LNR4), IIA (T1LNR3, T2LNR1-2, T3LNR1), IIB (T1LNR5, T2LNR3-4, T3LNR2, T4aLNR1), IIC (T2LNR5, T3LNR3-4, T4aLNR2, T4bLNR1), IIIA (T3LNR5, T4aLNR3-4, T4bLNR2), IIIB (T4aLNR5, T4bLNR3-4), and IIIC (T4bLNR5). In the training cohort, the TLNR had better prognostic discrimination [area under receiver operating characteristic curve (AUC), 0.621 vs. 0.608, P < 0.001], superior model-fitting ability in predicting overall survival [Akaike information criteria (AIC), 561,129 vs. 562,052], and better net benefits than the AJCC 8th tumor/node/metastasis (TNM) classification. Those results were successfully validated in predicting both overall and disease-free survivals in an independent validation set. Conclusions The TLNR classification has better prognostic discrimination, model-fitting ability and net benefits than the AJCC 8th TNM classification for better stratifying operable colon cancer patients, especially in patients with less than 12 retrieved lymph nodes.

A GAN-Based Video Intra Coding

Intra prediction is a vital part of the image/video coding framework, which is designed to remove spatial redundancy within a picture. Based on a set of predefined linear combinations, traditional intra prediction cannot cope with coding blocks with irregular textures. To tackle this drawback, in this article, we propose a Generative Adversarial Network (GAN)-based intra prediction approach to enhance intra prediction accuracy. Specifically, with the superior non-linear fitting ability, the well-trained generator of GAN acts as a mapping from the adjacent reconstructed signals to the prediction unit, implemented into both encoder and decoder. Simulation results show that for All-Intra configuration, our proposed algorithm achieves, on average, a 1.6% BD-rate cutback for luminance components compared with video coding reference software HM-16.15 and outperforms previous similar works.

Robustness and Accuracy Improvement of Data Processing with 2D Neutral Network for Transient Absorption Dynamics

To achieve the goal of efficiently analyzing the transient absorbance spectrum without arbitrary assumption and to overcome the limitation of conventional method in fitting ability and highly noised background, it...

ParamE: Regarding Neural Network Parameters as Relation Embeddings for Knowledge Graph Completion

We study the task of learning entity and relation embeddings in knowledge graphs for predicting missing links. Previous translational models on link prediction make use of translational properties but lack enough expressiveness, while the convolution neural network based model (ConvE) takes advantage of the great nonlinearity fitting ability of neural networks but overlooks translational properties. In this paper, we propose a new knowledge graph embedding model called ParamE which can utilize the two advantages together. In ParamE, head entity embeddings, relation embeddings and tail entity embeddings are regarded as the input, parameters and output of a neural network respectively. Since parameters in networks are effective in converting input to output, taking neural network parameters as relation embeddings makes ParamE much more expressive and translational. In addition, the entity and relation embeddings in ParamE are from feature space and parameter space respectively, which is in line with the essence that entities and relations are supposed to be mapped into two different spaces. We evaluate the performances of ParamE on standard FB15k-237 and WN18RR datasets, and experiments show ParamE can significantly outperform existing state-of-the-art models, such as ConvE, SACN, RotatE and D4-STE/Gumbel.