How to write a scientific paper?

Title of project:?Try not to use abbreviations — should be simple and informative

Abstract

Gives an overview - summary of the paper. Highlight the problem and give aim and goals, contributions - and results. How do you improve the body of knowledge? What are your major contributions? What makes your paper notable - what makes your work significant addition to the literature? Can your work be easily reproduced i.e - is the code/software - data source openly available?

(recommendation: use past/present tense and active verse)

1. Introduction

1. First paragraph: opening statement - eg. “ Neural networks along with deep learning models have gained a lot of popularity recently due to their success in multimedia applications [1][2] such as face recognition, biometrics, autonomous driving systems etc [3][4]." Ensure you cite review papers - old and new ones where possible when you are trying to summarize a field. Ensure that you cite the foundational papers, i.e which papers contributed to the field of neural networks, decision trees, ensemble methods, etc? (recommendation: use past tense and active verse)
2. Second paragraph: highlight some challenges of neural networks or time series problems (eg. big data problems, overfitting — bias-variance issue, hyperparameter tuning) (recommendation: use past tense and active verse)
3. Third paragraph: discuss other issues relating to your paper. Could be a relating issue bringing together multiple fields, cite your work where possible if you have already worked in the area and highlight the contributions. (recommendation: use past tense and active verse)
4. Fourth paragraph: motivation paragraph — why do you want to do this project? eg. not enough work was done that evaluates the key parameters for neural networks. Not enough work done that evaluates temporal CNNs for climate-related problems. (recommendation: use present tense and active verse)
5. Fifth paragraph — contribution paragraph: In this paper or project, we investigate (expand on the title of the project). Then list major goals (what would you test — hyperparameters? which ones ?) Algorithms? Which ones? What dataset would you use? Classificatifation problem or regression? (use present tense and active verse)
6. Overview: the rest of the paper is organised as follows. In section 2 …. In Section 3 … Finally, we conclude the paper in Section xxx

2. Background and Related Work

This section can have subsections that are related to your methodology or application problem, where you are covering major details that can't be given in the Introduction. This includes background information such as Adam training algorithm, basic details of MCMC, and basic details of the Genetic Algorithm. The aim of this section is to help general readers - this is useful in multidisciplinary papers so that readers find all information in one place. Also, you can review the latest literature in this area which is not really directly related but could be nearly distant to your project/paper. This helps in placing your paper in the spot - it shows where your neighbours are hanging out - well kind of ...

(recommendation: use past/present tense and active verse)

3. Methodology

1. Data - describe the original data (if application problem) and also the way to process the data. (recommendation: use present tense and active verse)
2. Overview of the methods (LSTM, CNN etc) (recommendation: use past tense and active verse)
3. Framework or workflow diagram — to highlight data input, processing, model, predictions and post-processing of results. (recommendation: use present tense and active verse)
4. Software suite — library (recommendation: use present tense and active verse)
5. Experiment setting where you give default settings such as neural network topology etc that won't be evaluated in experiments. The goal of this section is to ensure that others can reproduce what you have done - it is best to give GitHub repo link with this section or place the link end of the paper. (recommendation: use present tense and active verse)

4. Results

You need to layout the results — use present tense and active verse, and describe the results as if you are summarizing to someone who cannot see them.

eg. “In this section, we present the results of our proposed methodology. Table 1 presents the effect of the training time on the performance accuracy and Figure 2 shows the convergence plot. We notice a trend that the performance accuracy improves as the time increases but there comes a time (200 epochs) where there is not much improvement. This is not the case for the conventional method (Bayes-NN) where it takes more than 500 epochs to reach such a point.”

Follow these as examples of how results are described and discussed:

1. Chandra R; Saini R, 2021, ‘Biden vs Trump: Modeling US General Elections Using BERT Language Model’,?IEEE Access, vol. 9, pp. 128494–128505,?https://dx.doi.org/10.1109/ACCESS.2021.3111035
2. Chandra R; Goyal S; Gupta R, 2021, ‘Evaluation of Deep Learning Models for Multi-Step Ahead Time Series Prediction’,?IEEE Access, vol. 9, pp. 83105–83123,?https://dx.doi.org/10.1109/ACCESS.2021.3085085
3. Chandra R; Krishna A, 2021, ‘COVID-19 sentiment analysis via deep learning during the rise of novel cases’,?PLoS ONE, vol. 16, pp. e0255615,?https://dx.doi.org/10.1371/journal.pone.0255615
4. R. Chandra and V. Kulkarni, “Semantic and Sentiment Analysis of Selected Bhagavad Gita Translations Using BERT-Based Language Framework,” in?IEEE Access, vol. 10, pp. 21291–21315, 2022,?https://ieeexplore.ieee.org/document/9715095
5. Chandra R; He Y, 2021, ‘Bayesian neural networks for stock price forecasting before and during COVID-19 pandemic’,?PLoS ONE, vol. 16, pp. e0253217,?https://dx.doi.org/10.1371/journal.pone.0253217

Please note, that the above is only for guiding you —different fields have different requirements and styles, and you need to follow the requirements. (recommendation: use present tense and active verse)

5. Discussion

1. summarise your results — have you reached your goals?
2. implications of the results — can relate to the literature?
3. limitations of the methodology — results

(recommendation: use past tense and active verse)

6. Conclusions

1. what are the major contributions?
2. directions for future research

(recommendation: use past tense and active verse)

References

[1] Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feedforward networks are universal approximators.?Neural networks,?2(5), pp. 359–366. Retrieved from?https://www.sciencedirect.com/science/article/abs/pii/0893608089900208

[2] Trenn, S. (2008). Multilayer perceptrons: Approximation order and a necessary number of hidden units.?IEEE transactions on neural networks,?19(5), pp.836–844. Retrieved from?https://ieeexplore.ieee.org/document/4469950

APPENDIX

https://users.ics.aalto.fi/ntatti/howtowrite2016/tutorial.pdf

https://www.sciencedirect.com/science/article/pii/S1878764915001606

Youtube Video: