<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Sequence Models (RNNs &amp; LSTMs) on DATATWEETS</title><link>/courses/machine-learning/sequence-models/</link><description>Recent content in Sequence Models (RNNs &amp; LSTMs) on DATATWEETS</description><generator>Hugo</generator><language>en</language><copyright>Copyright (c) 2025 Datatweets</copyright><lastBuildDate>Fri, 14 Nov 2025 09:00:00 +0200</lastBuildDate><atom:link href="/courses/machine-learning/sequence-models/index.xml" rel="self" type="application/rss+xml"/><item><title>Lesson 1 - Introduction to Recurrent Neural Networks</title><link>/courses/machine-learning/sequence-models/lesson-1-introduction-to-recurrent-neural-networks/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-1-introduction-to-recurrent-neural-networks/</guid><description>Discover what sequence data is, why ordinary feedforward networks cannot model order, and how recurrent neural networks add a hidden state that carries memory forward through time. You will meet the running dataset for this module, the monthly S&amp;amp;P 500 index from 1950 to today, and load it with pandas and TensorFlow.</description></item><item><title>Lesson 2 - Basic RNN Architecture</title><link>/courses/machine-learning/sequence-models/lesson-2-basic-rnn-architecture/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-2-basic-rnn-architecture/</guid><description>Learn how a recurrent neural network actually works: the recurrence relation, unrolling through time, and backpropagation through time. Then build sliding windows from real S&amp;amp;P 500 data, scale them correctly, and train a Keras SimpleRNN to forecast the next month&amp;rsquo;s price.</description></item><item><title>Lesson 3 - Advanced RNN Architecture: GRU and LSTM</title><link>/courses/machine-learning/sequence-models/lesson-3-advanced-rnn-architecture-gru-and-lstm/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-3-advanced-rnn-architecture-gru-and-lstm/</guid><description>Learn why simple recurrent networks struggle to remember long-range patterns, meet the gated cells that solve it, and see the gate equations for LSTM and GRU. You will build SimpleRNN, GRU, and LSTM forecasters on the real S&amp;amp;P 500 monthly series and compare their test RMSE side by side.</description></item><item><title>Lesson 4 - Adding Convolutional Layers to RNNs</title><link>/courses/machine-learning/sequence-models/lesson-4-adding-convolutional-layers-to-rnns/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-4-adding-convolutional-layers-to-rnns/</guid><description>Discover how a 1D convolution slides a kernel along a sequence to detect local motifs, why causal padding keeps forecasting honest, and how to build a Conv1D + LSTM hybrid in Keras on real S&amp;amp;P 500 monthly data.</description></item><item><title>Lesson 5 - Time-Series Forecasting with RNNs</title><link>/courses/machine-learning/sequence-models/lesson-5-time-series-forecasting-with-rnns/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-5-time-series-forecasting-with-rnns/</guid><description>Turn a single price series into supervised learning pairs, split it the right way through time, scale it without leaking the future, and train an LSTM to forecast the S&amp;amp;P 500 one step ahead. You will evaluate the forecast with RMSE and MAE and learn why financial prediction is genuinely hard.</description></item><item><title>Lesson 6 - Guided Project: Forecasting the S&amp;P 500</title><link>/courses/machine-learning/sequence-models/lesson-6-guided-project-forecasting-the-sp-500/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>/courses/machine-learning/sequence-models/lesson-6-guided-project-forecasting-the-sp-500/</guid><description>Bring every piece of this module together in one guided project. You will load real monthly S&amp;amp;P 500 index data, build 12-month windows, split by time, scale, train an LSTM with Keras, evaluate it on a held-out test set, and reason carefully about what the model can and cannot do.</description></item></channel></rss>