Table 6.

Comparative overview of present work with existing ADLs monitoring and forecasting system.

TITLE	CATEGORY OF AAL APPLICATION	DATA COLLECTION	SET OF ACTIVITIES	TYPE OF MACHINE LEARNING ALGORITHM	RELIABILITY AND PERFORMANCE PARAMETERS, MERITS OF ANNOTATION, ACTIVITY DETECTION, AND FORECASTING
PROJECT
COBRA (CUMULATIVELY OVERLAPPING WINDOWING APPROACH FOR AMBIENT RECOGNITION OF ACTIVITIES) [37]	BEHAVIORAL RECOGNITION	USES CASAS DATASET. MORE THAN 6500 ACTIVITIES RECOGNIZED FROM IT.	SLEEPING, MEAL PREPARATION, RELAX, HOUSEKEEPING, EATING, LEAVE HOME, ENTER HOME, WORK	SLIDING WINDOW TECHNIQUE	RECOGNITION ACCURACY (0.821) AND RECALL (0.89)
AUTOMATED FEATURE ENGINEERING [38]	ACTIVITY DETECTION	OPEN SOURCE DATASETS OF DIFFERENT RESEARCH GROUPS WITH THE APPLICATION OF WEARABLE SENSORS	WALKING, STANDING, SITTING, VACUUMING, SWEEPING	SVM, NB, AND KNN	ACTIVITY DETECTION PRECISION (0.83) RECALL (0.80)
MINING HUMAN ACTIVITY PATTERNS FROM SMART HOME BIG DATA [39]	ACTIVITY DETECTION AND FORECASTING	400 MILLION SENSORS ACTIVATIONS	WATCHING TV, COOKING, USING COMPUTER, PREPARING FOOD AND CLEANING DISHES OR CLOTHES	BAYESIAN NETWORKS	PREDICTION ACCURACY (0.80)
LAPLACE [40]	ACTIVITY DETECTION AND FORECASTING	OPEN SOURCE DATASETS OF DIFFERENT RESEARCH GROUPS	WAKE UP, SHOWER, EAT, GOING OUT, RELAX, COOK	FREQUENT SEQUENTIAL PATTERN MINING	DID NOT mention THE PARAMETER VALUE, THE PERFORMANCE WAS AVERAGE
AGACY MONITORING [41]	ACTIVITY DETECTION	SENSING SYSTEM DEVELOPED	PREPARING FOOD, EAT, REST, DISHWASHING, WAKEUP	ONTOLOGICAL MODELING, SEMANTIC REASONING, AND DEMPSTER SHAFER THEORY	ACTIVITY DETECTION PRECISION (0.91) F1 SCORE (0.87) AND RECALL (0.83)
CONTEXTUALIZED BEHAVIOR PATTERNS [42]	BEHAVIORAL RECOGNITION AND FORECASTING	CASAS DATASETS OF 193 DAYS USED	MEAL PREPARATION, SLEEPING, WASH DISHES, WORK, ENTER HOME, LEAVE HOME, TOILET, HOUSEKEEPING, RELAX, EATING	CONTEXTUALIZED PREFIX-TREE	DID NOT claim ANNOTATION AND ADL RECOGNITION MERIT VALUES. THE FORECASTING PRECISION VALUES (0.392) AND RECALL VALUE (0.41)
AGING IN PLACE BY CHRITIAN DEBES [43]	BEHAVIORAL RECOGNITION	DATA WAS COLLECTED FROM TWO HOUSEHOLDS WITH MORE THAN 1000 ACTIVITY INSTANCES	PERSONAL HYGIENE, SLEEP WORK, MEAL PREPARATION, WATCH TV, SLEEP, SHOWERING	SVM, HMM AND FISHER KERNEL LEARNING (FKL)	THE ADL DETECTION CLASS AVERAGE ACCURACY FOR FKL (0.71), HMM (0.69) AND SVM (0.68)
ONLINE DAILY HABIT MODELING AND ANOMALY DETECTION (ODHMAD) MODEL [44]	BEHAVIORAL AND ANOMALY DETECTION	OBTRUSIVE AND UNOBTRUSIVE SENSING SYSTEM	MOVEMENT, OPEN-CLOSE STATES OF DOOR/WINDOW, FLUSH TOILET, USE OF ELECTRICAL DEVICES, TAKE SHOWER, WASH HAND, FALLS, EATING	ONLINE ACTIVITY RECOGNITION (OAR)	ANOMALY PRECISION (0.78), FALSE ALARM RATE (0.21) AND MISS DETECTION RATE (0.11)
WELLNESS INDEX MODEL	Behavioral PATTERN GENERATION AND ANOMALY DETECTION	UN-OBTRUSIVE HETEROGENEOUS WIRELESS SENSORS NETWORK	SLEEPING TAKING FOOD TOILET/LATRINE CALMING SHOWER/PERSONAL HYGIENE USE OF APPLIANCES	Novel Wellness Indices Modelling and Detection ALGORITHM	Sensitivity (0.9852), Specificity (0.9988), Precision (0.9887), Accuracy (0.9974), F1 score (0.9851), Correlation Coefficient (0.9144), False Negative Rate (0.0130)