Jayalath Ekanayake Jonas Tappolet Harald Gall Abraham Bernstein TECHNICAL REPORT No. IFI-2.4 Time variance and defect prediction in software projects: additional figures 2 University of Zurich Department of Informatics (IFI) Binzmühlestrasse 4, CH-85 Zürich, Switzerland ifi
Jayalath Ekanayake, Jonas Tappolet, Harald Gall, Abraham Bernstein Time variance and defect prediction in software projects: additional figures Technical Report No. IFI-2.4 Dynamic and Distributed Information Systems, Software Evolution and Architecture Lab Department of Informatics (IFI) University of Zurich Binzmuehlestrasse 4, CH-85 Zurich, Switzerland http://www.ifi.uzh.ch/ddis/, http://seal.ifi.uzh.ch
Time Variance and Defect Prediction in Software Projects. All Figures of Experiment 5. Figures..4 show the heat-maps of Eclipse, Mozilla, Netbeans and Open Office. These heatmaps show prediction quality variability when predicting bugs in one target period using different prediction models. The maximum, minimum, mean, and variance of the values as well as the histogram of variance s in each column of the heat-maps (Figures..4) are shown in Figures.5.8. Figures.9.2 show the variability in prediction quality when predicting bugs in different targets using the same model. The descriptive statistics of the values in each row of these heat-maps are shown in Figures.3.6.
2 Chapter. Time Variance and Defect Prediction in Software Projects 7 Mar6 5 6 Jul5 Number of months looked back from prediction 5 4 3 2 Aug3 Apr5 5 5 5 5 Jul May2 Mar3 Jan4 Nov4 Target prediction time Sep5 Jul6 May7 Figure.: Eclipse heat-map: Prediction quality on same target using different training periods with the point of highest highlighted
. All Figures of Experiment 5. 3 8 5 7 Number of months looked back from prediction 6 5 4 3 2 5 5 5 5 May Mar2 Jan3 Nov3 Sep4 Jul5 Target prediction time May6 Mar7 Jan8 Figure.2: Mozilla heat-map: Prediction quality on same target using different training periods with the point of highest highlighted
4 Chapter. Time Variance and Defect Prediction in Software Projects 7 5 6 Number of months looked back from prediction 5 4 3 2 5 5 5 5 Jul May2 Mar3 Jan4 Nov4 Target prediction time Sep5 Jul6 May7 Figure.3: Netbeans heat-map: Prediction quality on same target using different training periods with the point of highest highlighted
. All Figures of Experiment 5. 5 8 5 7 Number of months looked back from prediction 6 5 4 3 2 5 5 5 5 Jul May2 Mar3 Jan4 Nov4 Sep5 Target prediction time Jul6 May7 Mar8 Figure.4: Open Office heat-map: Prediction quality on same target using different training periods with the point of highest highlighted
6 Chapter. Time Variance and Defect Prediction in Software Projects May2 Jan4 Sep5 May7 May2 Jan4 Sep5 May7 (a) Maximum (b) Minimum x 3 7.5 5 2.5 May2 Jan4 Sep5 May7 May2 Jan4 Sep5 May7 (c) Mean (d) Variance 25 2 Frequency 5 5 2 3 4 5 6 7 8 Variance x 3 (e) Histogram: Variance Figure.5: Descriptive statistics of values in each column of the Eclipse heat-map (Figure.)
. All Figures of Experiment 5. 7 May Jan3 Sep4 May6 Jan8 May Jan3 Sep4 May6 Jan8 (a) Maximum (b) Minimum..75 variance.5.25 May Jan3 Sep4 May6 Jan8 May Jan3 Sep4 May6 Jan8 (c) Mean (d) Variance 5 4 Frequency 3 2..2.3.4.5.6.7.8.9. Variance (e) Histogram: Variance Figure.6: Descriptive statistics of values in each column of the Mozilla heat-map (Figure.2)
8 Chapter. Time Variance and Defect Prediction in Software Projects May2 Jan4 Sep5 May7 May2 Jan4 Sep5 May7 (a) Maximum (b) Minimum May2 Jan4 Sep5 May7 variance.3.275.25.225.2.75.5.25..75.5.25 May2 Jan4 Sep5 May7 (c) Mean (d) Variance 6 5 4 Frequency 3 2.5..5.2.25.3 Variance (e) Histogram: Variance Figure.7: Descriptive statistics of values in each column of the Netbeans heat-map (Figure.3)
. All Figures of Experiment 5. 9 Jul Mar3 Nov4 Jul6 Mar8 Jul Mar3 Nov4 Jul6 Mar8 (a) Maximum (b) Minimum..75 variance.5.25 Jul Mar3 Nov4 Jul6 Mar8 Jul Mar3 Nov4 Jul6 Mar8 (c) Mean (d) Variance 25 2 Frequency 5 5..2.3.4.5.6.7.8.9. Variance (e) Histogram: Variance Figure.8: Descriptive statistics of values in each column of the Open Office heat-map (Figure.4)
Chapter. Time Variance and Defect Prediction in Software Projects 69 5 59 5 Length of training period 49 39 5 29 5 9 5 9 Nov Sep2 Jul3 May4 Mar5 Jan6 Nov6 Figure.9: Eclipse heat-map: Prediction quality at different target periods
. All Figures of Experiment 5. 8 5 7 6 5 Length of training period 5 4 3 5 2 5 5 Jan Nov Sep2 Jul3 May4 Mar5 Jan6 Nov6 Sep7 Figure.: Mozilla heat-map: Prediction quality at different target periods
2 Chapter. Time Variance and Defect Prediction in Software Projects 7 5 6 5 5 Length of training period 4 3 5 2 5 5 Jan Nov Sep2 Jul3 May4 Mar5 Jan6 Nov6 Figure.: Netbeans heat-map: Prediction quality at different target periods
. All Figures of Experiment 5. 3 7 5 6 5 5 Length of training period 4 3 5 2 5 5 Jun Apr2 Feb3 Dec3 Oct4 Aug5 Jun6 Apr7 Figure.2: Open Office heat-map: Prediction quality at different target periods
4 Chapter. Time Variance and Defect Prediction in Software Projects 9 9 29 39 49 59 Length of training period 9 9 29 39 49 59 Length of training period in months (a) Maximum (b) Minimum..75.5.25 9 9 29 39 49 59 Length of training period 9 9 29 39 49 59 Length of training period (c) Mean (d) Variance 5 Frequency 5.2.4.6.8..2 Variance (e) Histogram: Variance Figure.3: Descriptive statistics of values in each row of the Eclipse heat-map (Figure.9)
. All Figures of Experiment 5. 5 2 3 4 5 6 7 8 (a) Maximum 2 3 4 5 6 7 8 (b) Minimum..75 variance.5.25 2 3 4 5 6 7 8 May Jan3 Sep4 May6 Jan8 (c) Mean (d) Variance 5 Frequency 5..2.3.4.5.6.7.8.9.. Variance (e) Histogram: Variance Figure.4: Descriptive statistics of values in each row of the Mozilla heat-map (Figure.)
6 Chapter. Time Variance and Defect Prediction in Software Projects 2 3 4 5 6 7 (a) Maximum 2 3 4 5 6 7 (b) Minimum..5 2 3 4 5 6 7 2 3 4 5 6 7 (c) Mean (d) Variance 4 2 Frequency 8 6 4 2..2.3.4.5.6.7.8.9. Variance (e) Histogram: Variance Figure.5: Descriptive statistics of values in each row of the Netbeans heat-map (Figure.)
. All Figures of Experiment 5. 7 7 7 27 37 47 57 67 77 (a) Maximum 7 7 27 37 47 57 67 77 (b) Minimum..5 7 7 27 37 47 57 67 77 7 7 27 37 47 57 67 77 (c) Mean (d) Variance 5 Frequency 5..2.3.4.5.6.7.8.9.. Variance (e) Histogram: Variance Figure.6: Descriptive statistics of values in each row of the Open Office heat-map (Figure.2)