An Activation Method of Topic Dictionary to Expand Training Data for Trend Rule Discovery

Algorithm 1

Pseudocode for learning trend rules.

(01) initializeTransaction(Tr)

(02) For 𝑑∈𝐷

(03)  𝑊_𝑑 = extractEvaluationObjectAndAttribute(𝑑);

(04)  𝑡_𝑑 =  getTime(𝑑);

(05)  initializeClassCounter(𝑁);

(06)  For 𝑤∈𝑊_𝑑

(07)   if 𝑤∈𝐸

(08)    𝑣_{𝑤,𝑡𝑑},𝑣_{𝑤,𝑡𝑑+1} = getNumericalValue𝑉[𝑤],𝑡_𝑑;

(09)    𝑐ℎ_𝑤 = calculateChangeRatio𝑣_{𝑤,𝑡𝑑},𝑣_{𝑤,𝑡𝑑+1};

(10)    𝑐_𝑤 = identifyClassForEvaluationObject𝑐ℎ_𝑤,𝑇ℎ;

(11)    𝑁𝑐_𝑤 + = 1;

(12)  𝑐_𝑑 = identifyClassForElement(𝑁);

(13)  if 𝑐_𝑑≠𝜙

(14)   Tr[𝑐_𝑑] = addTransaction𝑊_𝑑,𝑇𝑟[𝑐_𝑑];

(15) For 𝑐∈𝐶

(16)  𝑃[𝑐] = discoverFrequentPattern𝑇𝑟[𝑐],𝑀_𝑠;

(17)  𝑅[𝑐] = extractTrendRule(𝑃[𝑐],𝑐);