Stock Prices Prediction Using Machine Learning and Deep Learning Techniques (with Python codes)

Hi Want- AI/ML & big data-analytics programmer Area: Forex market trading Context: I have bot (EA) with anti-martingale strategy In Mt4 Trading Forex Platform in 1 min chart and doing profit in USA trading time 5:30 pm to 11:30 pm and given good profit in trendy market but not work in sideways market and other than USA market most of days. Explain the solution with application of ML ,AI and Big data analytics Solution should give two answer 1 Suitable time to trade (based on historical data and future forcast) ? 2. Suggest suitable trading strategy for a. trendy market b. sideways market 3. What will change in side- ways market get good amount of Profit? Any Question you can ask. Approach: 1. Data cleaning: Using basic signal processing algorithms to feed the cleanest input values 2. Data addition- Adding other relevant data if possible: Bid , Ask, Volume, Open, Close etc 3. Feature engineering 1- Using time series mathematical transformations like skew, deviation, lagged values, rolling window variables, time-based feature like vol_afternoon, mathematical transformation like logs etc. [login to view URL] known indicators to be fed into the algorithm like Bollinger bands, MFI, PCR etc 5. Feature reduction: PCA/t-SNE/UMAP to remove collinearity so that the Neural Network doesn’t get confused with too many dependent variables 6. Unsupervised learning to make clusters of the data before feeding it to the neural networks 7. Hyperparameter Tuning: Looping over hyperparameters to get best AUC values for best performance in production 8. Model Optimization: Looping around data techniques with hyperparameters to optimize the Neural Network 9. Algorithmic visualization: Using visualization tools in Python, we will try to makes sense of algorithm based on the practical understanding of the market and options considered 10. Going a layer deeper, we will segregate types and traits op options where out algorithm gives us best efficiencies to develop more trust in the algorithm