Algorithm 1 Contextual data extraction algorithm (CDEA)

Input: Sensor identities (SAID, SSID, ACTST, ACTET), where SAID is the main activity, SSID is the sub-activity identities, ACTST and ACTET are the activation and deactivation time of SAID (representing a main activity) respectively.

Output: Comprehensive training tuple (Time ^ Location ^ ACT ^ S_ACT ^ Dur_ACT → Class)

1: IF (SAID== ON state)

2:  IF (SAID == i)

3:     Location = Loci ^ Activity = ACTi ^ Object = Obji

4:  IF (SSID == j)

5:     Sub_Act = S_ACTj

6: IF (SAID == OFF state)

7:        Dur_ACTi = |ACTET − ACTST|

8:    ELSE  Dur_ACTi = undetermined

9:  Time = ACTST

10: RETURN Training tuple (TRD)