An unsupervised algorithm does either not have a training phase as a supervised algorithm does, or it has a training phase that uses unlabelled data (e.g. a Hopfield network). In both cases, the algorithm is itself responsible for discovering and modelling patterns inherent in the data that is presented to it.
- used by
- ALG_Adaptive resonance theory network ALG_Association rule learning ALG_Autoencoder ALG_Bayesian network ALG_DBSCAN ALG_Expectation maximization ALG_Factor analysis ALG_Hierarchical clustering ALG_Hopfield network ALG_Latent semantic indexing ALG_Local outlier factor ALG_Markov random field ALG_Multidimensional scaling ALG_Principal component analysis ALG_Probabilistic latent semantic indexing ALG_Projection pursuit ALG_Restricted Boltzmann machine ALG_Spherical k-means ALG_k-means ALG_k-medians ALG_k-medoids