RLS-A reduced labeled samples approach for streaming imbalanced data with concept drift

Extending battery life: A low-cost practical diagnostic technique for lithium-ion batteries

“…In such cases data mining techniques can be used to evaluate changes from healthy to faulty cells in order to maintain diagnosis performance on a BMS [33].…”

A reduced labeled samples (RLS) framework for classification of imbalanced concept-drifting streaming data.

“…The results are reported for the chunk size of 1000 and for the varied labeling percentage of (5,10,15,20,25,30,50,70,100). The map size used in our experiment is 10*10.…”

“…, AND LABELING PERCENTAGES (5,10,15,20,25,30,50,70,100 (400,600,800,1000,1200), AND LABELING PERCENTAGES (5,10,15,20,25,30,50,70,100 …”

“…CoverType data sets are shown in Figure 22 and Figure Adult data sets for chunk sizes (400,600,800,1000,1200), and labeling percentages (5,10,15,20,25,30,50,70,100) In Figure 22(a), the F-measure is reported for the weather data set. For this data set, at lower labeling such as 5%, the F-measure is also lower compared to 100% labeling for all the chunk sizes.…”

“…Table 6 shows the F-measure results of the Electricity data sets. The Required Labeling (5,10,15,20,25,30,50,70,100) are shown on the top of the table, and for each chunk size, the Real Labeling is shown as well. Table 6 demonstrates the Real labeling percentage is higher than Required labeling budget at lower chunk sizes, and it will be remained within the budget at higher chunk sizes.…”