HMM學(xué)習(xí)(3)-Patterns generated by a hidden process

2007-12-18 20:31 903人閱讀 matrixsystemreturningeachalgorithmsun

3. ?Patterns generated by a hidden process

3.2 When a Markov process may not be powerful enough

In some cases the patterns that we wish to find are not described sufficiently by a Markov process. Returning to the weather example, a hermit may perhaps not have access to direct weather observations, but does have a piece of seaweed. Folklore tells us that the state of the seaweed is probabilistically related to the state of the weather - the weather and seaweed states are closely linked. In this case we have two sets of states, the observable states (the state of the seaweed) and the hidden states (the state of the weather). We wish to devise an algorithm for the hermit to forecast weather from the seaweed and the Markov assumption without actually ever seeing the weather.

?

在一些場合我們所希望找到的模式不能夠被馬爾科夫過程很好的描述。回到天氣的例子，一個(gè)隱士可能沒有辦法直接觀察到天氣，但是他有一片海藻。傳說海藻與天氣的狀態(tài)有一定的聯(lián)系。在這個(gè)例子中，我們有兩個(gè)狀態(tài)集合，可觀察的狀態(tài)（海藻的狀態(tài)）和隱狀態(tài)（天氣的狀態(tài)）。我們希望為隱士設(shè)計(jì)一個(gè)算法，在不能實(shí)際看到天氣的情況下，來從海藻和馬爾科夫假設(shè)中預(yù)測出天氣的狀態(tài)。

A more realistic problem is that of recognizing speech; the sound that we hear is the product of the vocal chords, size of throat, position of tongue and several other things. Each of these factors interacts to produce the sound of a word, and the sounds that a speech recognition system detects are the changing sound generated from the internal physical changes in the person speaking.

一個(gè)更實(shí)際的問題是語音識(shí)別；我們所聽到的聲音是產(chǎn)生自聲帶，喉嚨的大小，舌頭的位置和其他一些東西。每一個(gè)因素相互作用產(chǎn)生了詞語的聲音，一個(gè)語音識(shí)別系統(tǒng)所探測到的聲音都是人說話時(shí)內(nèi)部身體變化所產(chǎn)生的變化的聲音。

Some speech recognition devices work by considering the internal speech production to be a sequence of hidden states, and the resulting sound to be a sequence of observable states generated by the speech process that at best approximates the true (hidden) states. In both examples it is important to note that the number of states in the hidden process and the number of observable states may be different. In a three state weather system (sunny, cloudy, rainy) it may be possible to observe four grades of seaweed dampness (dry, dryish, damp, soggy); pure speech may be described by (say) 80 phonemes, while a physical speech system may generate a number of distinguishable sounds that is either more or less than 80.

?

一些語音識(shí)別設(shè)備認(rèn)為內(nèi)部的語音的產(chǎn)物（internal speech production，語言？）是一個(gè)隱狀態(tài)的序列，發(fā)出的聲音是一個(gè)可觀察狀態(tài)的序列，這個(gè)序列由很好的近似了真實(shí)狀態(tài)（隱狀態(tài)）的語音過程所產(chǎn)生。在兩個(gè)例子中，非常重要的一點(diǎn)是，隱過程中的狀態(tài)的數(shù)量與可觀察狀態(tài)的數(shù)量會(huì)很不一樣。在3狀態(tài)的天氣系統(tǒng)中（天晴，多云，下雨），可能會(huì)觀察到海藻濕潤度的四個(gè)等級(jí)（干燥，稍干，微濕，潮濕）；純語音可以被80個(gè)音素所描述，而一個(gè)人體的語音系統(tǒng)可能會(huì)產(chǎn)生非常多的不同的聲音，比80多或少。

?

In such cases the observed sequence of states is probabilistically related to the hidden process. We model such processes using a hidden Markov model where there is an underlying hidden Markov process changing over time, and a set of observable states which are related somehow to the hidden states.

?

在這些情況下，狀態(tài)的可觀察序列在一定的概率下與隱過程相關(guān)聯(lián)。我們使用隱馬爾科夫模型來對這樣的過程進(jìn)行建模，這里有一個(gè)潛在的隱馬爾科夫過程隨時(shí)間而改變，以及一個(gè)在一定程度上與隱狀態(tài)關(guān)聯(lián)的可觀察的狀態(tài)集合。

3.2 Hidden Markov Models

The diagram below shows the hidden and observable states in the weather example. It is assumed that the hidden states (the true weather) are modeled by a simple first order Markov process, and so they are all connected to each other.

?

下圖展示了在天氣的例子中的隱狀態(tài)以及可觀察狀態(tài)。它假定了隱狀態(tài)（真實(shí)的天氣）通過一個(gè)簡單的一階馬爾科夫過程來進(jìn)行建模，所以他們兩兩之間都有連接。

?

?

The connections between the hidden states and the observable states represent the probability of generating a particular observed state given that the Markov process is in a particular hidden state. It should thus be clear that all probabilities `entering' an observable state will sum to 1, since in the above case it would be the sum of Pr(Obs|Sun), Pr(Obs|Cloud) and Pr(Obs|Rain).

?

在隱狀態(tài)和可觀察狀態(tài)之間的連接表示了在給定馬爾科夫過程停留在一個(gè)特定的隱狀態(tài)時(shí)產(chǎn)生特定的觀察狀態(tài)的概率。不難看出，所有的進(jìn)入一個(gè)可觀察狀態(tài)的概率之和應(yīng)該為1，在上面的例子中就應(yīng)該是Pr(Obs|Sun), Pr(Obs|Cloud) 和 Pr(Obs|Rain)三者的和。？？和底下的矩陣有什么區(qū)別？？

?

In addition to the probabilities defining the Markov process, we therefore have another matrix, termed the confusion matrix, which contains the probabilities of the observable states given a particular hidden state. For the weather example the confusion matrix might be;

?

除了定義了馬爾科夫過程的概率，我們還有另外一個(gè)矩陣，稱作混合矩陣（confusion matrix，先驗(yàn)概率？），它包括了給定特定隱狀態(tài)的情況下可觀察狀態(tài)的概率。天氣的混合矩陣可以是：

?

?

Notice that the sum of each matrix row is 1.

每行之和為1。

3.3 Summary

We have seen that there are some processes where an observed sequence is probabalistically related to an underlying Markov process. In such cases, the number of observable states may be different to the number of hidden states.

?

我們已經(jīng)看到在一些過程中，可觀察序列是在一定概率下與隱藏的馬爾科夫過程相關(guān)聯(lián)。在這些例子中，可觀察狀態(tài)的數(shù)量可以與隱狀態(tài)不同。

We model such cases using a hidden Markov model (HMM). This is a model containing two sets of states and three sets of probabilities;

我們使用隱馬爾科夫模型來對這些例子進(jìn)行建模。這個(gè)模型包含了兩個(gè)狀態(tài)集合和三個(gè)概率集（哪三個(gè)？轉(zhuǎn)移概率，confusion matrix，？初始？）。

• hidden states : the (TRUE) states of a system that may be described by a Markov process (e.g., the weather).
• observable states : the states of the process that are `visible' (e.g., seaweed dampness).

?

confusion matrix？？？？？？？？？？?

轉(zhuǎn)載于:https://www.cnblogs.com/hyubz/p/3620375.html

與50位技術(shù)專家面對面20年技術(shù)見證，附贈(zèng)技術(shù)全景圖

總結(jié)

以上是生活随笔為你收集整理的HMM学习(3)-Patterns generated by a hidden process的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。