摘要(英) |
In recent years, artificial intelligence has been rapidly developing, and with the progress of the internet, 5G is also widely used, and the metaverse is entering our lives. The foundation of this cannot be separated from deep learning, and TensorFlow is an important deep learning framework for language understanding tasks of machine learning. TensorFlow simplifies and improves the code library of DistBelief for computer scientists, making it faster and better, so that users can easily use the deep learning algorithms inside the code library to construct models. Nowadays, we widely use deep learning to analyze big data, among which, the Convolutional Neural Network (CNN) has good performance in image recognition. This article will introduce how to use Keras to construct a CNN model to analyze school student performance data. We will establish a model, when we have this student′s semester performance data, we can use this model to predict the student′s performance in the next semester.
After we acquire student performance data, it contains the grades and multiple fields of each student, these fields represent different meanings. Among them, we find out each student′s grades for each subject in each semester to analyze student performance. We use R language to process data, select college students′ data, and convert students′ performance data in each semester into images. We use students′ previous few semester grades to do linear regression to predict the grade for the next semester.
This article uses Keras to create a CNN model in R language, we find that the model′s parameter settings do not have a certain rule, the number of epochs is not the more the better and the learning rate is not the slower the better. We need to continuously train the model and adjust the model parameters to find the best model for processing student performance data.
In the early days, linear regression was commonly used for predicting values. Now that artificial intelligence is becoming more and more proficient, this article will explore whether the MLP and CNN models established using deep learning can have better results than traditional linear regression models, and the calculation of the machine will not pay too much the price.
|
參考文獻 |
1.Intel. How to Get Started as a Developer in AI.October 2016; Available from: https://software.intel.com/en-us/articles/how-to-get-started-as-a-developer-in-ai.
2.Gill, J.K. Automatic Log Analysis using Deep Learning and AI. 2018 [cited 2020 July 1]; Available from: https://www.xenonstack.com/blog/log-analytics-deep-machine-learning/.
3.McCulloch, W. S., & Pitts, W. (1943). A logical calculus of the ideas immanent in nervous activity. The bulletin of mathematical biophysics, 5(4), 115-133..
4.Sebastian Ruder (2016). An overview of gradient descent optimization algorithms.
5.LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
6.斎藤康毅, Deep Learning|用 Python 進行深度學習的基礎理論實作.2017
7.Keras.Why choose Keras,July 17,2020;From:https://keras.io/why_keras/
8.https://keras.rstudio.com/reference/optimizer_adam.html
9.林大貴,TensorFlow+Keras深度學習人工智慧實務應用,2017:碩博文化
10.Yann LeCun, Yoshua Bengio & Geoffrey Hinton, Deep learning, Nature, Vol. 521, pp. 436–444 , 2015.
11.Y. Lecun, L. Bottou, Y. Bengio and P. Haffner, "Gradient-based learning applied to document recognition," in Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, Nov. 1998, doi: 10.1109/5.726791.
12.Kingma, D. P., & Ba, J. L. (2015). Adam: a Method for Stochastic Optimization. International Conference on Learning Representations
13.David A. Freedman (2009). Statistical Models: Theory and Practice. Cambridge University Press. p. 26. A simple regression equation has on the right hand side an intercept and an explanatory variable with a slope coefficient. A multiple regression e right hand side, each with its own slope coefficient
14.Hardesty, Larry (14 April 2017). "Explained: Neural networks". MIT News Office. Retrieved 2 June 2022.
15.Teh, Y.W. and Hinton, G. E. Rate-coded restricted boltzmann machines for face recognition. In Advances in Neural Information Processing Systems, volume 13, 2001.
16.Agarap, Abien Fred (2018). Deep Learning using Rectified Linear Units (ReLU)
17.Valueva, M.V.; Nagornov, N.N.; Lyakhov, P.A.; Valuev, G.V.; Chervyakov, N.I. (2020). "Application of the residue number system to reduce hardware costs of the convolutional neural network implementation". Mathematics and Computers in Simulation. Elsevier BV. 177: 232–243. doi:10.1016/j.matcom.2020.04.031. ISSN 0378-4754. S2CID 218955622. Convolutional neural networks are a promising tool for solving the problem of pattern recognition.
18.Zhang, Wei (1988). "Shift-invariant pattern recognition neural network and its optical architecture". Proceedings of Annual Conference of the Japan Society of Applied Physics.
|