1. Introduction

Road accident is a special case of trauma that constitutes a major cause of disability and untimely death. It has been estimated that over 300,000 persons die and 10 to 15 million persons are injured every year in road accidents throughout the world. Statistics have also shown that mortality in road accidents is very high among young adults that constitute the major part of the work force. In actual fact, accidents kill faster than AIDS and it gives no preparatory time to its victims. In order to combat this problem, various road safety strategies have been proposed and used. These methods mainly involve conscious planning, design and operations on roads. One important feature of this method is the identification and treatment of accident prone locations commonly called black spots; black spots are not the only cause of accidents on the highway. Also various organizations such as Police High Way Patrol, Vehicle Inspection Officer (VIO), Federal Road Safety Commission (FRSC) among others are charged with the responsibility of maintaining safety thereby reducing road accidents. However, lack of good forecasting techniques has been a major hindrance to these organizations in achieving their objectives.

It is against this background that Decision Tree is beingproposed to model data from road accident database to determine causes of accidents and accident prone locations using historical data collected from Ibadan-Lagos express road as reference point.

2. Objective

The primary objective of this research is to use data mining technique; decision tree to predict causes of accident and accident prone locations on highways using data collected on Lagos – Ibadan express way.

3. Methods

3.1. Data Mining

Data Mining is an interactive process of discovering valid and novel, useful and understandable patterns or models in large database (Han, Mannila and Smyth, 2001). Data Mining, according to Han, Mannila and Symth (2001) is a process that uses a variety of data analysis tools to discover patterns and relationships in data that may be used to make a valid prediction. Data mining uses advances in the field of Artificial Intelligence (AI) and Statistical techniques. Therefore, decision tree is being used in this research

3.2. Decision Trees

Decision Trees have emerged as a powerful technique for modelling general input / output relationships. They are tree – shaped structures that represents a series of roles that lead to sets of decisions. They generate rules for the classification of a dataset and a logical model represented as a binary (two – way split) tree that shows how the value of a target variable can be predicted by using the values of a set predictor variables. Decision trees, which are considered in a regression analysis problem, are called regression trees. Thus, the decision tree represents a logic model of regularities of the researched phenomenon.

3.3. Accidents along Lagos - Ibadan Express Way

Lagos to Ibadan Express road is one of the busiest roads in Africa. This is because. Lagos was the capital of Nigeria until the seat of government moved to the Federal Capital Territory Abuja and also the headquarters of many national institutions while Ibadan is said to be the largest city in black Africa. The traffic along this route is very heavy because it is a gateway linkage of the heavy traffic going from the Northern, Eastern and Majority of Western states. Fig 3.1 shows the frequency of accidents between the distances of 1 and 40km from Ibadan to Lagos between January 2002 and December 2003.The statistics shows that having a means of predicting likely location of accident base on some input values is essential to advice on dangerous locations.

Figure 3.1. Graph of Frequency of Accidents against Month

Several works have been carried out by different researchers both on road accident analysis and forecasting, using Decision Tree and Artificial Neural Networks. Martin, Grandal and Pilkey (2000), analysed the relationship between road infrastructure and safety by using a cross-sectional time-series data base collected for all 50 U.S. states over 14 years. The result suggested that as highway facilities are upgraded, there are reduced fatalities. Gelfand (1991) studied the effect of new pavement on traffic safety in Sweden. The result of his study shows that Traffic accidents increased by 12 % after one year of resurfacing on all types of roads. Akomolafe (2004) employed Artificial Neural Network using multilayer perceptron to predict likelihood of accident happening at particular location between the first 40 kilometers along Lagos-Ibadan Express road and discovered that location 2 recorded the highest number of road accident occurrence and that, tyre burst was the major cause of accident along the route. Ossenbruggen (2005) used a logistic regression model to identify statistically significant factors that predict the probabilities of crashes and injury crashes aiming at using these models to perform a risk assessment of a given region. Their study illustrated that village sites are less hazardous than residential and shopping sites. Abdalla et al (1987) studied the relationship between casualty frequencies and the distance of the accidents from the zones of residence. As might have been anticipated, the casualty frequencies were higher nearer to the zones of residence, possibly due to higher exposure. Akomolafe et al (2009) used geo spatial technology to identify various positions along major roads in Nigeria. The study revealed that the casualty rates amongst residents from areas classified as relatively deprived were significantly higher than those from relatively affluent areas.

Table 3.1. Record of Accidents along Lagos Ibadan between year 2002 and 2003 S/NO Month No of Accident 1 Jan 2002 6 2 Feb 2002 11 3 March 2002 10 4 April 2002 18 5 May 2002 14 6 June 2002 4 7 July 2002 6 8 August 2002 1 9 September 2002 9 10 October 2002 6 11 Nov. 2002 4 12 December 2002 5 13 Jan 2002 5 14 Feb 2003 5 15 March 2003 4 16 April 2003 7 17 May 2003 2 18 June 2003 1 19 July 2003 4 20 August 2003 5 21 September 2003 8 22 October 2003 5 23 Nov. 2003 5 24 December 2003 6

3.4. Process of Data Mining

The process of data mining consists of three steps which are:

3.4.1. Data Preparation

This includes; Data collection, Data cleaning and Data transformation.

3.4.2. Data Modeling

This research considers the data of accident record between the first 40km from Ibadan to Lagos. The data were organized into a relational database.

The unknown causes in Table 3.2 may include other factors such as Law enforcement agent problems, attitude of other road users, inadequate traffic road signs, traffic congestion and general vehicle conditions

The sample data used covered the period of 24 Months, that is, January 2002 to December 2003 as indicated in Fig. 3.1.

The output variable is the location and the locations can be divided into three distinct regions tagged regions A, B and C, meaning we have three outputs. Where

First location 1 – 10km is Region A or location 1, Above10km – 20km is region B or Location 2 and above 20km is region C or Location 3

The data sample used covered a period of twenty four Months starting from January 2002 to December 2003.The data were collected by Akomolafe (2004) and this is presented in Table.3.3.

3.4.3. Deployment

In this stage, new sets are applied to the model selected in the previous stage to generate predictions or estimates of the expected outcome.

Table 3.2. showing variables given both continuous and categorical values

Table 3.3. Sample Data collected from FRSC (Akomolafe O.P 2004) SNO DATE TYPE TIME SEASON CAUSE LOCATION REG. NO 1 6.1.2002 2 2 1 2 31 XG 506 LND 2 7.1.2002 2 1 1 1 14 XC 720 ACD 3 11.1.2002 1 1 1 1 14 AM 713 LND 4 12.1.2002 2 1 1 2 27 XE 905 JJJ 5 19.1.2002 1 2 1 3 27 AA 559 LAF 6 30.01.02 3 3 1 2 12 AA 156 NWD 7 03.02.02 2 2 1 2 35 XF 635 JJJ 8 05.02.02 2 1 1 2 10 XE 141 AKD 9 05.02.02 2 3 1 2 14 XE 124 AKD 10 06.02.02 2 3 1 2 31 XE 124 AKD 11 11.02.02 1 1 1 3 5 AG 276 LAR 12 14.02.02 1 1 1 2 14 13 18.02.02 1 2 1 2 18 14 21.02.02 2 1 1 2 19 XD 249 SMK 15 21.02.02 3 2 1 2 19 XC 361 KTU 16 24.02.02 2 1 1 2 18 XE 716 SMK 17 27.02.02 2 3 1 2 35 XC 307 SGM 18 03.03.02 2 1 1 2 16 XE 807 NSR 19 05.03.02 1 2 1 2 10 XC 348 AKP 20 07.03.02 2 1 1 2 2 OY 2270 JB 21 07.03.02 1 1 1 2 13 AP 820 LSD 22 07.03.02 3 2 1 2 18 XE 322 APP 23 19.03.02 2 2 1 2 19 XC 993 AGL 24 19.03.02 3 1 2 2 LA 1804 RF 25 30.03.02 1 4 1 2 14 AM 343 FST 26 31.03.02 1 2 1 2 14 KC 461 ABA 27 31.03.02 1 2 1 2 14 BS 142 KJA 28 01.04.02 2 1 2 2 22 AA 807 EGB 29 01.04.02 1 1 2 2 22 BX 527 GGE 30 01.04.02 2 2 2 2 18 AG 787 GNN 31 02.04.02 1 1 2 1 7 AU 725 MAP 32 02.04.02 2 2 2 2 27 XG 358 APP 33 04.04.02 1 1 2 2 15 CY 65 EKY 34 04.04.02 1 2 2 2 17 AJ 21 AGG 35 05.04.02 1 2 2 1 6 AW 45 FST 36 06.04.02 2 1 2 2 30 XB 855 AKD 37 07.04.02 1 2 2 1 13 AL 567 YAB 38 09.04.02 2 2 2 2 12.5 XA 787 WWP 39 13.04.02 2 1 2 1 1 XB 791 GNN 40 13.04.02 2 1 2 1 11 XA 127 AFN 41 13.04.02 1,2 1 2 1 11 AH 202 AKN 42 22.04.02 1 2 2 1 15 RA 01 KRD 43 22.04.02 1,3 2 2 1 11 BB 731 KJA 44 27.04.02 2 2 2 2 27 AU 739 JJJ 45 28.04.02 1 2 2 1 14 AE 316 FST 46 03.04.02 1 3 2 2,1 12 AZ 824 AAA 47 5.8.2002 1 1 2 2 20 AA 654 GBY 48 5.8.2002 1 1 2 2 30 XF 65 JJJ 49 5.10.2002 2&1 1 2 1 35 DM 207 AAA 2 BL 86 AAA 50 5.10.2002 1 1 2 1&2 35 BR 608 LSR 51 5.11.2002 3 1 2 2 26 XB 606 APP 52 5.13.2002 2 1 2 1 2 XA 616 YLW 53 5.13.2002 1 1 2 1 26.5 BM 566 GGE 54 5.14.2002 2 3 2 2 15 XC 348 AKD 55 5.15.2002 1 2 2 2 19 OY 2077 JB 56 5.15.2002 1 2 2 2 14 AJ 101 NND 57 5.20.2002 1 2 2 26 AU 682 ABC 58 5.21.2002 2 2 2 24 XG 719 FST 59 5.25.2002 1 1 2 2 12 AV 70 LSR 60 6.2.2002 3 2 1 12 AZ 191 MUS 61 6.3.2002 2 2 2 2 16 AQ 742 YYY 62 6.15.2002 2 1 2 2 12 XA 682 YRE 63 6.16.2002 1 1 2 2 21 AL 885 AKN 64 6.16.2002 2 1 2 2 21 XE 751 SMK 65 7.15.2002 2 1 2 3 12 XH 649 GGE 66 7.20.2002 2 2 2 2 10 XB 286 KNR 67 8.8.2002 3 2 2 2 12 XE 232 SGM 68 9.19.2002 1 3 2 2 22 XA 940 KNH 69 9.20.2002 2 1 2 2 4 AX 94 JJJ 70 9.20.2002 3 2 2 2 7 XC 768 BDJ 71 9.21.2002 1 1 2 1 29 BL 254 SMK 72 9.21.2002 2 1 2 1 16 AP 647 AKR 73 9.21.2002 2 2 2 18 XC 253 GGE 74 9.22.2002 2 1 2 2 10 LA 979 BG 75 9.22.2002 2 3 2 2 16 XU 510 GGE 76 9.27.2002 2 2 2 12 77 10.1.2002 1 2 2 1 6 AA 05 MHA 78 10.14.2002 2 1 2 2 13 XE 869 MUS 79 10.16.2002 2 2 2 2 15 XB 888 AKR 80 10.29.2002 2 2 2 7 81 10.29.2002 2 2 2 2 17 XD 168 BDJ 82 10.29.2002 3 1 2 2 6 AA 342 LES 83 11.4.2002 2 1 1 1 5 BX 877 KJA 84 11.10.2002 2 1 1 2 12 XC 637 RKJ 85 11.10.2002 2 2 1 2 11 XC 937 SGM 86 11.12.2002 1 1 12 AA 466 KNR 87 2.12.2004 2 1 1 2 14 XG 182 JJJ 88 12.7.2002 3 2 1 2 1 XA 425 CRC 89 12.10.2002 2 3 1 3 13 XD 695 EKY 90 12.11.2002 2 2 1 2 16 XA 350 EDY 91 12.12.2002 1 1 2 14 XG 955 KSF 92 23.01.2002 1 3 1 1 16 XA 411 EJG 93 18.01.03 1 3 1 1 18 AE 015 GBN 94 27.01.03 2 2 1 2 8 XD 125 LSR 95 29.01.03 3 4 1 2 12 XC 616 KTU 96 29.01.03 2 1 2 14 XF 797 AKD 97 02.02.03 2 1 1 2 18 CW 293 AAA 98 12.02.03 1 2 1 1 18 AV 3 GGE 99 12.02.03 2 2 1 2 18 XB 6 WWD 100 12.02.03 1 3 1 1 12 HB 40 KJA 101 17.02.03 2 3 1 2 11 XB 446 MNY 102 05.03.03 1 2 1 2 6 AE 753 KRE 103 19.03.03 2 1 1 2 12 XH 382 ABC 104 28.03.03 3 1 1 12 AG 145 NRK 105 31.03.03 2 3 1 2 13 AA 499 GBY 106 05.04.03 2 2 2 3 11.5 XD 432 KSF 107 06.04.031 1 1 2 3 12 CE 188 JJJ 108 06.04.03 2 1 2 2 12 FA 01 JJ 109 14.04.03 1 1 2 28 FV 43 AAA 110 24.04.03 1 2 2 2 7 OY 01 SE 111 24.04.03 3 2 2 2 9 XB 328 MAG 112 30.04.03 3 3 2 1 16 XD 644 NRK 113 10.05.03 1 2 40 AA 399 KTU 114 16.05.03 1 3 2 2 20 XH 327 ADC 115 02.06.03 1 1 2 1 8 XB 144 YRE 116 20.07.03 2 1 2 2 27 5K 324 LND 117 26.07.03 1 2 2 2 9 DG 329 LSR 118 28.07.03 2 2 2 2 13 XJ 179 LND 119 28.07.03 2 2 2 1 18 XF 114 EPE 120 02.08.03 1 1 2 2 13 CB 434 MUS 121 02.08.03 1 1 2 1 8 XG 954 FST 122 09.08.03 1 1 2 1 19 AG 802 SGB 123 16.08.03 2 2 2 2 2 XF 450 SMK 124 31.08.03 1 1 2 1 14 OY 1281 TD 125 01.09.03 3 2 2 1 8 XA 362 KJA 126 08.09.03 1 2 18 XH 723 JJJ 127 14.09.03 2 19 128 16.09.03 1 2 2 2 6 AA 112 YRE 129 21.09.03 2 1 2 2 31 XB 766 AGG 130 24.09.03 2 2 2 1 18 XC 115 EDE 131 28.09.03 2 1 2 2 14 XN 739 AAA 132 28.09.03 2 3 2 2 13 XD 642 NRK 133 06.10.03 1 2 2 2 11 DG 548 LND 134 14.10.03 2 2 2 2 12 XA 730 FUF 135 18.10.03 2 3 2 2 28 XA 286 GBH 136 19.10.03 1 2 2 22 AA 188 AAA 137 20.10.03 2 2 2 2 27 LG 016 KNE 138 01.11.03 3 1 1 1 9 XA 847 KEH 139 02.11.03 2 2 1 2 18 XC 575 GGE 140 25.11.03 1 1 1 3 24 BO 984 APP 141 27.11.03 1 1 1 2 18 AJ 06 SGB 142 27.11.03 2 2 1 2 13 XB 369 EKY 143 06.12.03 2 1 1 2 13 AP 938 KJA 144 09.12.03 3 3 1 1 13 BM 130 MAP 145 13.12.03 2 1 1 1 7 XA 610 ARP 146 22.12.03 1 1 1 1 11 BL 500 GGE 147 24.12.03 1 1 3 12 JB 356 KJA 148 24.12.03 2 2 1 2 13 XG 562 AKD

4. Results

4.1. Analysis

The major step required to obtain result of the research was carried out by analysing the data using WEKA. WEKA is a collection of machine learning algorithms and data processing tools. It contains various tools for data pre-processing, classification, regression, clustering, association rules and visualization. There are many learning algorithms implemented in WEKA including Bayesian classifier, Trees, Rules, Functions, Lazy classifiers and miscellaneous classifiers. The algorithms can be applied directly to a data set. WEKA is also data mining software developed in JAVA it has a GUI chooser from which any one of the four major WEKA applications can be selected. For the purpose of this study, the Explorer application was used.

The Explorer window of WEKA has six tabs. The first tab is pre- process that enables the formatted data to be loaded into WEKA environment. Once the data has been loaded, the preprocess panel shows a variety of information as shown in figure 4.3 below.

Figure 4.1. WEKA GUI chooser

Figure 4.2. WEKA Explorer

4.1. Weka Classifiers

There are several classifiers available in WEKA but Function Tree and Id3 were used in this study in case of Decision Tree. Prism Rule based learner was generated using WEKA. Attribute importance analysis was carried out to rank the attribute by significance using information gain. Finally, correlation based feature subset selection (cfs) and consistency subset selection (COE) filter algorithm were used to rank and select the attribute that are most useful. The F- measure and the AUC which are well known measures of probability tree learning was used as evaluation metrics for model generated by WEKA classifiers.

Several numbers of setups of decision tree algorithms have been experimented and the best result obtained is reported as the data set. Each class was trained with entropy of fit measure, the prior class probabilities parameter was set to equal, the stopping option for pruning was misclassification error, the minimum n per node was set to 5, the fraction of objects was 0.05, the maximum number of nodes was 100, surrogates was 5, 10 fold cross-validation was used, and generated comprehensive results.

The best decision tree result was obtained with Id3 with 115 correctly classified instances and 33 incorrectly classified instances which represents 77.70% and 22.29% respectively.

Mean absolute error was 0.1835 and Root mean squared error was 0.3029.

The tree and rules generated with Id3 algorithm are given thus:

4.2. Id3 Tree

TYREBURST = TRUE

| SEASON = WET

| | TYPE = HAEVY VEHICLE

| | | TIME = EVENING: LOCATION2

| | | TIME = AFTERNOON: LOCATION2

| | | TIME = MORNING: LOCATION2

| | | TIME = NIGHT: null

| | TYPE = SMALL CAR: LOCATION2

| | TYPE = MOTOCYCLE: null

| SEASON = DRY

| | TIME = EVENING

| | | TYPE = HAEVY VEHICLE: LOCATION2

| | | TYPE = SMALL CAR: LOCATION3

| | | TYPE = MOTOCYCLE: null

| | TIME = AFTERNOON

| | | TYPE = HAEVY VEHICLE: LOCATION2

| | | TYPE = SMALL CAR: LOCATION2

| | | TYPE = MOTOCYCLE: null

| | TIME = MORNING

| | | TYPE = HAEVY VEHICLE: LOCATION3

| | | TYPE = SMALL CAR: LOCATION3

| | | TYPE = MOTOCYCLE: null

| | TIME = NIGHT: null

TYREBURST = FALSE

| TIME = EVENING

| | OVERSPEEDING = FALSE: LOCATION2

| | OVERSPEEDING = TRUE

| | | TYPE = HAEVY VEHICLE: LOCATION2

| | | TYPE = SMALL CAR: LOCATION2

| | | TYPE = MOTOCYCLE: null

| TIME = AFTERNOON

| | LOSS-OF-CONTROL = FALSE

| | | OVERSPEEDING = FALSE

| | | | BRAKE-FAILURE = FALSE

| | | | | TYPE = HAEVY VEHICLE

| | | | | | WRONG-OVERTAKING = FALSE

| | | | | | | BROKEN-SHAFT = FALSE: LOCATION1

| | | | | | | BROKEN-SHAFT = TRUE: LOCATION3

| | | | | | WRONG-OVERTAKING = TRUE: LOCATION2

| | | | | TYPE = SMALL CAR

| | | | | | SEASON = WET: LOCATION3

| | | | | | SEASON = DRY

| | | | | | | CARELESSDRIVING = FALSE: LOCATION3

| | | | | | | CARELESSDRIVING = TRUE: LOCATION2

| | | | | TYPE = MOTOCYCLE: LOCATION3

| | | | BRAKE-FAILURE = TRUE

| | | | | TYPE = HAEVY VEHICLE: LOCATION1

| | | | | TYPE = SMALL CAR: LOCATION1

| | | | | TYPE = MOTOCYCLE: LOCATION2

| | | OVERSPEEDING = TRUE

| | | | TYPE = HAEVY VEHICLE: LOCATION2

| | | | TYPE = SMALL CAR

| | | | | SEASON = WET: LOCATION2

| | | | | SEASON = DRY: LOCATION2

| | | | TYPE = MOTOCYCLE: null

| | LOSS-OF-CONTROL = TRUE

| | | TYPE = HAEVY VEHICLE: LOCATION2

| | | TYPE = SMALL CAR

| | | | SEASON = WET: LOCATION2

| | | | SEASON = DRY: LOCATION1

| | | TYPE = MOTOCYCLE: LOCATION1

| TIME = MORNING

| | SEASON = WET

| | | OVERSPEEDING = FALSE

| | | | TYPE = HAEVY VEHICLE

| | | | | WRONG-OVERTAKING = FALSE

| | | | | | CARELESSDRIVING = FALSE: LOCATION1

| | | | | | CARELESSDRIVING = TRUE: LOCATION2

| | | | | WRONG-OVERTAKING = TRUE: LOCATION1

| | | | TYPE = SMALL CAR

| | | | | CARELESSDRIVING = FALSE

| | | | | | LOSS-OF-CONTROL = FALSE: LOCATION3

| | | | | | LOSS-OF-CONTROL = TRUE: LOCATION2

| | | | | CARELESSDRIVING = TRUE: LOCATION1

| | | | TYPE = MOTOCYCLE: LOCATION2

| | | OVERSPEEDING = TRUE: LOCATION2

| | SEASON = DRY

| | | BROKEN-SHAFT = FALSE

| | | | TYPE = HAEVY VEHICLE

| | | | | CARELESSDRIVING = FALSE

| | | | | | LOSS-OF-CONTROL = FALSE

| | | | | | | BROKEN-SPRING = FALSE

| | | | | | | | OVERSPEEDING = FALSE: LOCATION2

| | | | | | | | OVERSPEEDING = TRUE: LOCATION2

| | | | | | | BROKEN-SPRING = TRUE: LOCATION2

| | | | | | LOSS-OF-CONTROL = TRUE: LOCATION2

| | | | | CARELESSDRIVING = TRUE: LOCATION3

| | | | TYPE = SMALL CAR

| | | | | CARELESSDRIVING = FALSE

| | | | | | OVERSPEEDING = FALSE

| | | | | | | UNKNOWN-CAUSES = FALSE

| | | | | | | | ROBBERY-ATTACK = FALSE

| | | | | | | | | WRONG-OVERTAKING = FALSE

| | | | | | | | | | LOSS-OF-CONTROL = FALSE

| | | | | | | | | | | TREE-OBSTRUCTION = FALSE

| | | | | | | | | | | | BRAKE-FAILURE = FALSE: LOCATION3

| | | | | | | | | | | | BRAKE-FAILURE = TRUE: LOCATION2

| | | | | | | | | | | TREE-OBSTRUCTION = TRUE: LOCATION2

| | | | | | | | | | LOSS-OF-CONTROL = TRUE: LOCATION2

| | | | | | | | | WRONG-OVERTAKING = TRUE: LOCATION2

| | | | | | | | ROBBERY-ATTACK = TRUE: LOCATION3

| | | | | | | UNKNOWN-CAUSES = TRUE: LOCATION3

| | | | | | OVERSPEEDING = TRUE: LOCATION3

| | | | | CARELESSDRIVING = TRUE: LOCATION1

| | | | TYPE = MOTOCYCLE: null

| | | BROKEN-SHAFT = TRUE: LOCATION3

| TIME = NIGHT: LOCATION2Prism rules

----------

Rule 1 If BROKEN-SHAFT = TRUE then LOCATION3

Rule 2 If ROBBERY-ATTACK = TRUE

and TYPE = SMALL CAR then LOCATION3

Rule 3 If TREE-OBSTRUCTION = TRUE

and TIME = EVENING then LOCATION3

Rule 4 If TYREBURST = TRUE

and TIME = MORNING

and TYPE = SMALL CAR

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 5 If TYPE = MOTOCYCLE

and CARELESSDRIVING = TRUE then LOCATION3

Rule 6 If ROAD-PROBLEM = TRUE

and TYPE = SMALL CAR

and TIME = AFTERNOON

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 7 If TYREBURST = TRUE

and SEASON = DRY

and TIME = MORNING

and TYPE = HAEVY VEHICLE

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 8 If UNKNOWN-CAUSES = TRUE

and TYPE = SMALL CAR

and TIME = MORNING

and SEASON = DRY then LOCATION3

Rule 9 If TYREBURST = TRUE

and TYPE = HAEVY VEHICLE

and TIME = AFTERNOON

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 10 If TIME = MORNING

and OVERSPEEDING = TRUE

and TYPE = SMALL CAR

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 11 If TYREBURST = TRUE

and TIME = EVENING

and TYPE = SMALL CAR then LOCATION3

Rule 12 If TYREBURST = TRUE

and TYPE = HAEVY VEHICLE

and TIME = AFTERNOON

and SEASON = WET

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 13 If TIME = MORNING

and LOSS-OF-CONTROL = TRUE

and TYPE = HAEVY VEHICLE

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 14 If UNKNOWN-CAUSES = TRUE

and TYPE = SMALL CAR

and TIME = MORNING

and SEASON = WET

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 15 If TYREBURST = TRUE

and TYPE = HAEVY VEHICLE

and SEASON = WET

and TIME = EVENING

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 16 If TIME = MORNING

and TYREBURST = TRUE

and TYPE = HAEVY VEHICLE

and SEASON = WET

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION3

Rule 17 If CARELESSDRIVING = TRUE

and TYPE = HAEVY VEHICLE

and SEASON = DRY then LOCATION3

Rule 18 If TIME = MORNING

and TYPE = SMALL CAR

and SEASON = DRY

and CARELESSDRIVING = FALSE

and WRONG-OVERTAKING = FALSE

and LOSS-OF-CONTROL = FALSE

and TREE-OBSTRUCTION = FALSE

and BRAKE-FAILURE = FALSE then LOCATION3

Rule 19 If TIME = NIGHT then LOCATION2

Rule 20 If WRONG-OVERTAKING = TRUE

and TYPE = SMALL CAR then LOCATION2

Rule 21 If TIME = EVENING

and CARELESSDRIVING = TRUE then LOCATION2

Rule 22 If TIME = EVENING

and UNKNOWN-CAUSES = TRUE then LOCATION2

Rule 23 If TIME = EVENING

and LOSS-OF-CONTROL = TRUE then LOCATION2

Rule 24 If TIME = EVENING

and ROBBERY-ATTACK = TRUE then LOCATION2

Rule 25 If TIME = EVENING

and TYPE = HAEVY VEHICLE

and SEASON = DRY then LOCATION2

Rule 26 If SEASON = WET

and TYPE = MOTOCYCLE then LOCATION2

Rule 27 If SEASON = WET

and OVERSPEEDING = TRUE

and TIME = MORNING then LOCATION2

Rule 28 If TYREBURST = TRUE

and SEASON = WET

and TYPE = SMALL CAR then LOCATION2

Rule 29 If TYREBURST = TRUE

and SEASON = WET

and TIME = MORNING

and TYPE = HAEVY VEHICLE

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 30 If TYPE = HAEVY VEHICLE

and ROBBERY-ATTACK = TRUE then LOCATION2

Rule 31 If TYPE = HAEVY VEHICLE

and OVERSPEEDING = TRUE

and TIME = AFTERNOON then LOCATION2

Rule 32 If TYREBURST = TRUE

and SEASON = WET

and TIME = EVENING

and TYPE = HAEVY VEHICLE

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 33 If TYREBURST = TRUE

and SEASON = WET

and TYPE = HAEVY VEHICLE

and TIME = AFTERNOON

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 34 If TYPE = HAEVY VEHICLE

and TIME = EVENING then LOCATION2

Rule 35 If TYPE = HAEVY VEHICLE

and OVERSPEEDING = TRUE

and TIME = MORNING

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 36 If TYREBURST = TRUE

and TIME = AFTERNOON

and TYPE = SMALL CAR

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 37 If BRAKE-FAILURE = TRUE

and TYPE = MOTOCYCLE then LOCATION2

Rule 38 If WRONG-OVERTAKING = TRUE

and TIME = AFTERNOON then LOCATION2

Rule 39 If TREE-OBSTRUCTION = TRUE

and TIME = MORNING then LOCATION2

Rule 40 If BROKEN-SPRING = TRUE

and TYPE = HAEVY VEHICLE

and TIME = MORNING

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 41 If TYPE = HAEVY VEHICLE

and TYREBURST = TRUE

and TIME = AFTERNOON

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 42 If LOSS-OF-CONTROL = TRUE

and TIME = MORNING

and TYPE = SMALL CAR then LOCATION2

Rule 43 If UNKNOWN-CAUSES = TRUE

and TYPE = HAEVY VEHICLE

and SEASON = DRY then LOCATION2

Rule 44 If OVERSPEEDING = TRUE

and TIME = AFTERNOON

and SEASON = WET then LOCATION2

Rule 45 If TYPE = HAEVY VEHICLE

and LOSS-OF-CONTROL = TRUE

and TIME = MORNING

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 46 If SEASON = WET

and LOSS-OF-CONTROL = TRUE

and TIME = AFTERNOON

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE

and TYPE = HAEVY VEHICLE then LOCATION2

Rule 47 If CARELESSDRIVING = TRUE

and TIME = AFTERNOON

and TYPE = SMALL CAR then LOCATION2

Rule 48 If OVERSPEEDING = TRUE

and TIME = AFTERNOON

and TYPE = SMALL CAR

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 49 If SEASON = WET

and TIME = EVENING

and TYPE = SMALL CAR

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and LOSS-OF-CONTROL = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = TRUE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

Rule 50 If TYPE = HEAVY VEHICLE

and LOSS-OF-CONTROL = TRUE

and TIME = AFTERNOON

and SEASON = DRY

and WRONG-OVERTAKING = FALSE

and CARELESSDRIVING = FALSE

and TYREBURST = FALSE

and OVERSPEEDING = FALSE

and TREE-OBSTRUCTION = FALSE

and PUSHED-BY-A-CAR = FALSE

and BROKEN-SHAFT = FALSE

and BROKEN-SPRING = FALSE

and BRAKE-FAILURE = FALSE

and ROAD-PROBLEM = FALSE

and UNKNOWN-CAUSES = FALSE

and ROBBERY-ATTACK = FALSE then LOCATION2

5. Discussion

There are 50 rules generated from this tree. Rule 1- 18 indicate the occurrence of accident in Location 3 and rule 19-50 also shows the occurrence of accident in location 2.This indicate that, location 2 has the highest number of road accident occurrence with Heavy-vehicle in the afternoon and during the dry season.

Rule 41 is the best one that can be used for prediction. The rule says that, Tyre bust is the cause of road accident with heavy vehicle within location 2 in the day time and during the dry season.Decision Tree Performance Analysis on Id3

Table 5.1. Detailed Accuracy By class Class TP rate FT rate Precision Recall F- measure Roc Area Location (3) 0.688 0.069 0.733 0.688 0.71 0.942 Location (2) 0.897 0.361 0.78 0.897 0.834 0.888 Location (1) 0.517 0.025 0.833 0.517 0.638 0.95 Weighted Avg. 0.777 0.232 0.78 0.777 0.769 0.912

Table 5.2. Confusion matrix Predicted category Actual category Location (3) Location (2) Location (1) Location (3) 22 10 0 Location (2) 6 78 3 Location (1) 2 12 15

Decision Tree performance Analysis on Function Tree (FT)

Table 5.3. Detailed Accuracy by Class Class TP rate FT rate Precision Recall F- measure Roc Area Location (3) 0.625 0.086 0.667 0.625 0.645 0.869 Location (2) 0.77 0.361 0.753 0.77 0.761 0.736 Location (1) 0.586 0.101 0.586 0.586 0.586 0.832 Weighted Avg. 0.703 0.25 0.702 0.703 0.702 0.783

Table 5.4. Confusion Matrix Predicted category Actual category Location (3) Location (2) Location (1) Location (3) 20 12 0 Location (2) 8 67 12 Location (1) 2 10 17

6. Conclusions

Using WEKA software to analyze accident data collected on Lagos-Ibadan road, it was found that decision tree can accurately predict the cause(s) of accident and accident prone locations along the road and other roads if relevant data are gathered and analyzed as in this case.

In Decision Tree Performance analysis, the, dataset were experimented with two algorithms; Id3 and FT (function tree) For Id3 algorithm, there were 115 correctly classified instances and 33 incorrectly classified instances which represent 77.70% and 22.29% respectively. Mean absolute error was 0.1835 and Root mean squared error was 0.3029.

Also for functional tree algorithm (FT), total number of tree size was 5 with 105 correctly classified instances representing 70.27% and 44 incorrectly classified instances representing 29.73%.

From the detailed accuracy by class and confusion matrix, Id3 attained accuracy rate of 0.777 and FT attained accuracy rate of 0.703.