Software
Products
My Programs and Technology
My Programs and Technology
Traider is a complex conglomeration of Neural Networks designed to acquire historic financial data from the foreign exchange and run machine learning algorithms on the data. This allows Traider to make intelligent predictions regarding future market activity based on pattern recognition in the forex market. Whilst this product is in its early stages, it performs relatively well.
The service is capable of receiving up to 100K tokens of pre-train data before it begins to assess realtime data. The system is also capable of fully automated trading with pre-defined limit behaviours for stoploss assignement and takeprofit placement. Balance percentage is implemented to allow for a balance related trading behaviour.
A dedicated financial prediction project built on neural networks and historic data to predict the increase/decrease of foreign exchange rates.
Making use of in-built machine learning and custom designed neural networks to provide predictions with an indefinitely increasing degree of accuracy. The software boasts a pre-train token set of up to 100k tokens.
Operation
Dedicated financial trend analysis network. Built to identify numerical patterns on the foreign exchange market (forex). An in-built self training algorithm allows the neural networks to be pre-trained with a range of 0 - 100K tokens of historic financial data.
Open Source as always.
Open for criticism/construction/contribution.
The content of this repository is available on (GITHUB) and can be cloned for free with no copyright laws.
The project is constantly updated and new logic and programming will be performed on the system to improve it over time. The project will be accompanied by a formal paper and remain under continual work and development through the forseeable future. Language conversion can be expected from python towards rust or C.
NET 19 is a neural network developed to make future predictions from a designated dataset. The neural network makes use of adaptive intelligence, also known as machine learning, to improve the accuracy of the model in real time. Programmed in python due to powerful machine learning capabilities.
Exceedingly powerful neural network. Dedicated to data profiling, NET 19 provides a powerful dataset and adaptive capabilities.
Trained on UK Gov data for traffic behaviour and vehicle criminality, NET 19, can identify probability of criminal offences from a few simple factors.
Weightings
The weightings the network operates on are designated into 2 sections.
The network uses "Categories" for predictions. Each category has a designated weighting value.
Within each category there are predefined "solutions". Each solution also has a designated weighting. The weighting of a category should be a decimal value < 1.
The sum of all category weightings is == 1. The value for each category weighting is defined by the adaptive intelligence from the solutions within.
Likewise the weighting of a category solution should be a decimal value < 1.
The sum of all solutions weightings within a category == 1. The value of any category solution weighting is defined by adaptive intelligence.
Calculation
When calculating the percentage chance, the magnitude of the category weighting is determined by the selected solution's weighting. E.g, a solution weighting of 0.3 means that the category weighting used in final calculation is only 30% of the total value.
The content of this repository is available on (GITHUB) and can be cloned for free with no copyright laws.
The project is constantly updated and new datasets and content will be attributed to the repository when suitable. The project will be accompanied by a formal paper and remain under continual work and development through the forseeable future. Language conversion can be expected from python towards rust or C.
A project that makes use of a collection of powerful custom neural networks working in conjunction to take news stories from media sites, summarise the stories, and perform image manipulation to create captivating news stories. This project is open source and is packaged with a readme file for easy usage. The development of this project spanned over a 3 day period and is capable of providing vast quantities of news stories carefully manipulated.
Languages Used: Python, JS, NodeJS
This service provides a powerful user interface and generation network to create graphics and literature from news headlines and summaries.
1. The network makes use of a range of News API's to find live stories relating to the supplied user prompt.
2. The stories are analysed using a neural network to identify the primary Headline of the news story.
3. The neural network reads the story description from the news API and summarises the story with the range of networks.
4. The image provided is then manipulated to place the story headline on the image in the appropriate place.
5. Story and image are evaluated and prepared for confirmation for the user.
Coming Soon.
The project is in early alpha stages and will be released to my github repository soon. A User Interface is implemented with a display window and the ability to alter generation parameters for optimal content generation.
When working through a series of computational projects, often byproducts of the process are vast quantities of computational function scripts, these can be imported into any project to improve performance or aid the development process. This repository contains a dump of program files available for import and usage in any project of your pleasing.
If you wish to use any of these scripts in your own projects, feel free to clone this git repository.
The software available in this github repository follow a few basic design principles.
All the software is written in the following stack: Python, HTML, CSS, JavaScript, NodeJS, PHP, Bash, or SQL.
I have used Objecy Orientated Programming in most cases and will continue to do so for the benefit of easy implementation and fast modification.
The python programs have detailed function descriptions set for easy editing with VSCode and other IDE's.
The software includes README files where suitable.
Most programs are written in the latest edition of each language and include a requirements.txt file for easy installation.
All programs are written in English and are copyright free, however some credit would be much appreciated.
The content of this repository is available on (GITHUB) and can be cloned for free with no copyright laws.
The project is constantly updated and new content is attributed to the repository when suitable. The project is expected to receive maintenance until late 2025 due to the continual working addition of python programming, however plans for conversion towards more Rust and C based programs is inevitable.