本頁只刊出中文翻譯與中文說明；英文原文請見下方原文連結。
原文連結
論文資訊
- 類型：已發表論文
- 日期：2017-05-25
摘要
Scientific explanation often requires inferring maximally predictive features from a given data set. Unfortunately, the collection of minimal maximally predictive features for most 隨機 processes is uncountably infinite. In such cases, one compromises and instead seeks nearly maximally predictive features. Here, we derive upper bounds on the rates at which the number and the coding cost of nearly maximally predictive features scale with desired predictive power. The rates are determined by the 碎形 dimensions of a process' mixed-state distribution. These results, in turn, show how widely used finite-order 馬可夫 models can fail as predictors and that mixed-state predictive features can offer a substantial improvement.
※ 此為已發表論文，全文需透過期刊付費取得