Neural Network Trading Bot

Understand how http requests establish client-server communication through a request–response cycle. Learn get and post methods, url parameters, status codes, headers, and the optional message body.

Explore how an application programming interface enables machines and users to interact with servers, using a waiter analogy, get and post requests, and JSON responses.

Learn how to read the API documentation, identify the base endpoint and JSON data format, and implement signed requests with public and private keys for trading actions.

Explore how api endpoints work, compare get and post requests, and learn to pass mandatory parameters like symbol and interval to fetch time and candlestick data.

Install postman to send requests to the Nancy API, download from getpostman.com, install on your operating system, and explore headers and response output for neural network trading bot workflows.

Master using Postman to send get requests to a financial API, test endpoints, and interpret responses, including server time, exchange data, and candlesticks with symbol and interval.

Develops a board order trading block that fetches cryptocurrency data, analyzes trends, and issues buy or sell orders through the audit and order engines.

Initialize the trading board by loading public and private keys from a credentials file and preparing a key-based options object for secure cryptocurrency trades.

Extract price details from Binance by building and sending API requests. Parse k-lines into Python objects and format open, close, high, low, and volume data for trading insights.

Learn to implement a signed request by hashing a private key with a secret, generating a timestamped hex signature with sha-256, and sending a request with an api key.

Set up a python logger for the neural network trading bot, creating a file handler, defining format and level, and logging credentials status and API request errors.

Construct market buy and sell orders by specifying symbol and trading amount, with optional arguments. Format the quantity to eight decimals and submit a signed post request to trading API.

Implement an infinite loop that fetches candlesticks, scales up or down candles for reduced error, and prints execution IDs before pausing between iterations.

Explore how the simple moving average serves as a key financial indicator, filtering price noise to reveal short-term and long-term market trends, with Python pandas demonstrations.

Explore how the exponential moving average weights recent prices more through a decay factor to smooth price data while preserving short-term fluctuations, enabling long-term stock or cryptocurrency trend analysis.

Explore the macd oscillator built from 12, 26, and 9 period exponential moving averages, calculated as (ema12 - ema26 + ema9)/3, to gauge price direction in stocks and crypto.

Identify the current problems in a modular trading board for multiple exchanges, including varying request formats, and the trade-off between modularity and performance.

Learn how Gekko serves as a flexible trading board that connects to exchanges, imports data, selects neural net strategies, and performs back tests to simulate profits before live trading.

Install Gecko on your local machine, clone the repository, install dependencies with npm, and run the Gecko UI to backtest strategies and explore neural-network driven trading.

Build a neural network trading strategy with a JavaScript library, train via stochastic gradient descent, and export the neural net digest strategy for Gekko.

Describe the strategy functions of a neural network trading bot, including init, learn with stochastic gradient descent, set normalize factor, update, on trade, predict candle, check, end, and logging.

Initialize the neural network by setting the name and history, apply smoothing, define input, hidden, and output layers, and add a stop-loss indicator with a threshold.

Explore the basic building block of neural networks: the neuron, which takes multiple inputs, processes them through a central nucleus, and yields corresponding outputs that connect to other neurons.

Learn how neuron inputs are weighted and combined: assign weights between 0 and 1, sum weighted inputs, and determine firing by a threshold of 1 to produce 1 or 0.

Explore how biases and weights influence neural network decisions by adding to input sums, shaping neuron firing, and producing outputs between 0 and 1 with activation functions.

Explore activation functions in neural networks, understanding forward propagation, unit and nonlinear activations, including the sigmoid and linear units, then apply by building an xor gate to test the model.

Implement the learn function by looping through price candles, using a price buffer and volumes, training the neural network with forward propagation, and updating the prediction count for all prices.

Normalize neural network inputs by converting large numbers to a 0 to 1 range using a scale of 10^digits, then divide inputs before feeding them to the network.

Implement the on data function to process buy and hold actions, initialize long orders, manage stop loss indicators, and log actions and prices to train the neural network predicting candles.

Implement a loss function for a neural network trading bot, using minimum predictions, computing mean vs current price deviation, and generating buy or sell orders with stop-loss and thresholds.

Process candles with an update function that uses a moving average, normalizes data into a price buffer, runs five training cycles, and logs stop loss events.

Extend the neural network trading bot beyond cryptocurrencies to stocks and commodities by using API endpoints for prices, candles, and orders, and train on time-based data to cover multiple assets.

Explore how a multi-nodal input neural network predicts prices by using past prices, market volume, moving averages, and 52-week highs and lows to improve forecasting.