Sparse matrix factorization with <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" altimg="si9.svg"><mml:msub><mml:mi>L</mml:mi><mml:mrow><mml:mn>2</mml:mn><mml:mo>,</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:msub></mml:math> norm for matrix completion

Jin, Xiao-Bo; Miao, Jianyu; Wang, Qiufeng; Geng, Guanggang; Huang, Kaizhu

doi:10.1016/j.patcog.2022.108655

Cited by 8 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From the forecasting results, it can be seen that RTNMFFM(0) can also obtain good results in some cases. This model based on the

norm measure of matrix factorization error is also known as indirect sparse matrix factorization with

norm [ 33 ], which indirectly optimizes the upper bound of the F-norm function in a way that is the reason for its ability to find more effective information from sparse data and obtain more accurate predictions. In order to compare the variability of the predictions, we also compared the results of the Diebold–Mariano test [ 45 ] for RTNMFFM and the best alternative model, and the conclusions are shown in Appendix B .…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Robust Multi-Dimensional Time Series Forecasting

Shen,

He,

Qin

2024

Entropy

View full text Add to dashboard Cite

Large-scale and high-dimensional time series data are widely generated in modern applications such as intelligent transportation and environmental monitoring. However, such data contains much noise, outliers, and missing values due to interference during measurement or transmission. Directly forecasting such types of data (i.e., anomalous data) can be extremely challenging. The traditional method to deal with anomalies is to cut out the time series with anomalous value entries or replace the data. Both methods may lose important knowledge from the original data. In this paper, we propose a multidimensional time series forecasting framework that can better handle anomalous values: the robust temporal nonnegative matrix factorization forecasting model (RTNMFFM) for multi-dimensional time series. RTNMFFM integrates the autoregressive regularizer into nonnegative matrix factorization (NMF) with the application of the L2,1 norm in NMF. This approach improves robustness and alleviates overfitting compared to standard methods. In addition, to improve the accuracy of model forecasts on severely missing data, we propose a periodic smoothing penalty that keeps the sparse time slices as close as possible to the time slice with high confidence. Finally, we train the model using the alternating gradient descent algorithm. Numerous experiments demonstrate that RTNMFFM provides better robustness and better prediction accuracy.

show abstract

“…From the forecasting results, it can be seen that RTNMFFM(0) can also obtain good results in some cases. This model based on the

norm measure of matrix factorization error is also known as indirect sparse matrix factorization with

Section: Methodsmentioning

confidence: 99%

“…Too many parameters in the model can also cause overfitting. To the best of our knowledge, the

-norm has better robustness properties [ 33 , 34 ], and non-negative matrix factorization can reduce overfitting.…”

Section: Related Workmentioning

confidence: 99%

Robust Multi-Dimensional Time Series Forecasting

Shen,

He,

Qin

2024

Entropy

View full text Add to dashboard Cite

show abstract

“…Recently, however, it was found to be applied to the prediction of student grades. For example, the problem of the prediction of students' curriculum grades in the next semester is analogized to the problem of score prediction or that of the next basket recommendation [23][24][25]. The method of the recommendation system has brought a new perspective to solve the problem of student grades prediction.…”

Section: Deep Matrix Factorizationmentioning

confidence: 99%

Study on Score Prediction Model with High Efficiency Based on Deep Learning

Yang

Bai

2022

Electronics

View full text Add to dashboard Cite

In the problem of unified classroom performance prediction, there is a certain lag in the prediction, and there are also problems such as the data sparsity and single feature in the data. In addition, feature engineering is often carried out manually in modeling, which highly depends on the professional knowledge and experience of engineers and affects the accuracy of the prediction to a certain extent. To solve the abovementioned gaps, we proposed an online course score prediction model with a high time efficiency that combines multiple features. The model uses a deep neural network, which can automatically carry out feature engineering and reduce the intervention of artificial feature engineering, thus significantly improving the time efficiency. Secondly, the model uses a factorization machine and two kinds of neural networks to consider the influence of first-order features, second-order features, and higher-order features at the same time, and it fully learns the relationship between the features and scores, which improves the prediction effect of the model compared to using only single feature learning. The performance of the model is evaluated on the learning analysis dataset from Fall 2015 to Spring 2021 and includes 412 courses with 600 students. The experimental results show that the performance of the prediction model based on the feature combination proposed in the present study is better than the previous performance prediction model. More importantly, our model has the best time efficiency of below 0.3 compared to the other models.

show abstract