Device-Oriented Automatic Semantic Annotation in IoT

Algorithm 2

Annotation dictionary matching algorithm.

Input:

An annotation dictionary , a word frequency dictionary TF and a new non-function description .

Output:

A property division result , which contains three components, i.e., Pref, Inter and WorkCond. Those three components

are the contents about the performance, interface and working condition of our device description framework.

Step 1. Obtain a word sequence after segment .

Step 2. For each in :

If   in , the category that belongs to.

(i) find the position of in and , marked as and .

(ii) IF   has more than one, choose a which can maximize .

Else  .

Then obtain a position sequence .

Step 3. For each in :

(i) If , If  , add to the component of that belongs to.

(ii) If , add to .

(iii) If  , add to .

(iv) If  , add to .

Return: