Latest notes
Understanding an Example of LDL Decomposition
Take the original matrix: A = | 4 2 | | 2 5 |. The following steps can be used to compute the LDL decomposition. 1. Initialize a unit lower-triangular matrix L and a diagonal matrix D: L = | 1 0 | | l 1 |, D = | d1 0 | | 0 d2 |. 2. By the definition of the LDL decomposition we have A = L * D * L^T. We need suitable values of l and d that satisfy this equation. First compute the first elements of L and D: 4 = 1 * d1 * 1 +
Data AnalysisSome Simple Thoughts on Neural Networks
Put simply, a model is really just a converter — or rather, it plays the role of a complex function that transforms input data into another form of presentation. If the transformed form is intuitive enough for the corresponding task, that task becomes easy to accomplish. Looking at things now, meta-learning has diverged somewhat from the original vision: early meta-learning aimed to let machines learn how to learn, but that was only a concept — in practice it has not truly been achieved. In ordinary transfer learning, one usually trains a model for a task first, then
Reflections on Using SVMs with Small Samples
Today when validating test data, I found that when using cross_val_score for cross-validation, the larger the proportion of the randomly allocated test set, the higher the accuracy—and the smaller the test set, the more likely the average accuracy is to be low. There are two major pitfalls when deriving models with SVM. One is that after train_test_split, when the data volume is insufficient, the learned sample distributions differ greatly and are extremely unstable; how to find the optimal train_test_split is hard to determine. The other is whether the assessment of the model’s generalization ability is reliable—and that is bound up with the dataset partitioned by train_test_split. This directly means that occasionally scoring high on the test set is merely an illusion...
3 posts