Machine Learning Approach Towards Road Accident Analysis In India

Abstract

This paper is an attempt to analyse the performance of four major machine-learning paradigms applied to modelling the occurrence of traffic accidents. The methods considered are support vector machines, decision trees, K-nearest neighbour and logistic regression. For the greatest possible accident reduction effects with limited budgetary resources, it is important that measures be based on scientific and objective surveys of the causes of accidents and severity of injuries.

1. Introduction

India is one of the developing countries, where the rate of road crashes is more than the critical limit. The road transportation increases year by year, but the rate of road crashes also increases with it. Road accidents are a human tragedy, which involve high human suffering. They impose a huge socio-economical cost in terms of untimely deaths, injuries and loss of potential income. The ramification of road accidents can be colossal, and its negative impact is felt not only on individuals, their health and welfare, but also on the economy. Consequently, road safety has become an issue of national concern. The huge number of injury and death due to road traffic accident reveals the story of global crisis of road safety. Road collisions are the second leading cause of death for people between the ages of 5 and 29 and third leading cause for people between 30 and 44. With the number of vehicles rapidly rising in developing countries, this epidemic is quickly worsening in low and middle-income countries and is on its way to becoming the third leading cause of death and disabilities by the year 2020 (WHO 2000). The loss in road traffic accidents enormous in economy and health related issues. Families having accident victims shatters with death and the victims seriously injured often needs medical facilities for the rest of their life and eventually becomes a burden to their family. Road traffic injuries are burdening health care systems in countries around the world. Patterns involved in dangerous crashes could be detected if we develop accurate prediction models capable of automatic classification of type of injury severity of various traffic accidents. These behavioural and roadway accident patterns can be useful to develop traffic safety control policies. Experiment results reveal that among the machine learning paradigms considered.

1.1 Objectives

The main objective of the road accident prediction system:

• Analyse previously occurred accidents in the locality, which will help us determine the most accident-prone area and help us set up the immediate required help for them.
• To minimise occurrence of accident while analysing severity of previously occurring accidents and taking into consideration other factors like pollution, visibility, alcohol content, age, weather etc.

1.2 Problem Statement

There are several problems with current practices for prevention of the accidents occurred in the localities. The database we will use is available officially by many institutes and government websites. The data collected will be analysed, integrated and grouped together based on different constraints using the best-suited algorithm. This estimation will be helpful to analyse and identify the flaw and the reasons of the accidents. It will also be helpful while making roads and bridges as a reference to avoid the same problems faced before. The predictions made will be very much useful to plan the management of such problems

2. Technology used

Machine learning in essence, is the concept of making a computer program that improves the performance by automatically learning and adapting with experience. It is basically working on many hypothesis, and then finding the best one that fits the observed data. It makes computers modify their action so that these actions get more and more accurate with experience, where accuracy is measured by how well the chosen action reflect the correct ones. Machine learning is a growing technology used to mine knowledge from data (known as data mining). Wherever data exists, things can be learned from it. Whenever there is excess of data, the mechanics of learning must be automatic. Machine learning technology is meant for automatic learning from voluminous datasets. Machine learning algorithms can be supervised or unsupervised. A data scientist or analyst works on supervised algorithms, and in unsupervised learning, we just have an input data. The various applications of machine learning are Web Search Engine, Photo Tagging Applications, Span Detectors, etc.

3. Related Work

In the previous years, many research papers have been published on these. Athanasios Theofilatos et al. [1] in 2014 presented A Review of The Effect of Traffic and Weather Characteristics on Road Safety. Despite the existence of generally mixed evidence on the effect of traffic parameters, a few patterns can be observed. For instance, traffic flow seems to have a nonlinear relationship with accident rates, even though some studies suggest a linear relationship with accidents. Hoang Nguyen et al. [2] in 2018 determined a model on automatic classification of traffic incident’s severity using machine learning approaches. The NSW TMC (Transport Management Centre) and the research organization Data61 in Sydney have collaborated to discover and visualize frequent patterns in historical incident response records, leading to the automatic classification of severity levels among past incidents using advanced machine learning, active learning and outlier detection techniques. Rakesh Mehar et al. [3] in 2013 highlighted the deficiencies in the present state of the art and also presents some basic concepts so that systematic approach for formulation of a road safety improvement program in India can be developed. Meenu Gupta et al. [4] in 2018 presented a Data Mining Approach of Accident Occurrences Identification with Effective Methodology and Implementation. Here, data mining is used to identify a new cause for at tan effect across the globe. It identifies the Accident Occurrences in different regions and to identify the most valid reason for happening accidents over the globe. Caliendo, C., De Guglielmo et al. [5] in 2019 provided an analysis of the frequency of total accidents (accidents involving material damage, physical injuries and fatalities), which occurred in 226 unidirectional motorway tunnels over a four-year monitoring period, based on unrelated and correlated random-parameter models. The so-called random-intercept model, in which only the regression intercept is assumed to be random, was also developed a priori for recording the random-effects (temporal correlations among accidents occurring in the same tunnel in different years). V. A. Olutayo, A. A. Eludire et al. [8] in 2014 employed Artificial Neural Networks and Decision Trees data analysis techniques to discover new knowledge from historical data about accidents in one of Nigeria’s busiest roads in order to reduce carnage on our highways. Lamija Herceg et al. [10] in 2019 extracted useful information from two road accident datasets, sourced from Fatality Analysis Reporting System in the United States of America, using data mining techniques. Naïve Bayes and C4.5 Decision Tree techniques were applied to form several predictions.

4. Study Area and Data Set

Data sets considered in this paper are taken from government websites and nationally authenticated sources that have undergone serious vetting and been authenticated. This ensures that the raw data taken for analysis is of minimum error hence, making our predictions as accurate as possible.

5. Methodology

Models created using accident data records, which can help to understand the characteristics of many features like driver’s behaviour, roadway conditions, light condition, weather conditions and so on. This can help the users to compute the safety measures, which is useful to avoid accidents. It can be illustrated how statistical method based on directed graphs, by comparing two scenarios based on out-of-sample forecasts. The model is performed to identify statistically significant factors, which can be able to predict the probabilities of crashes, and injury that can be used to perform a risk factor and reduce it.

Here the road accident study is done by analysing some data by giving some queries which is relevant to the study. The queries like what is the most dangerous time to drive, what fractions of accidents occur in rural, urban and other areas. What is the trend in the number of accidents that occur each year, do accidents in high speed limit areas have more casualties and so on. These data can be accessed using Microsoft excel sheet and the required answer can be obtained. This analysis aims to highlight the data of the most importance in a road traffic accident and allow predictions to be made. The results from this methodology can be seen in the next section of the report.

6. Implemented Work

Figure 1.1: Types of Vehicles

Figure 1.2: Age Range

Figure 1.3: No. of Accidents in a Month

Figure 1.4: Type of Roads

Figure 1.5: Light Condition

7. Conclusion

Road Accidents are caused by various factors. It can be concluded that Road Accident cases are hugely affected by the factors such as types of vehicles, age of the driver, age of the vehicle, weather condition, road structure and so on. Thus, we have built an application which gives efficient prediction of road accidents based on the above-mentioned factors.

References

1. Theofilatos, A., & Yannis, G. (2014). A review of the effect of traffic and weather characteristics on road safety. Accident Analysis & Prevention, 72, 244-256.
2. Nguyen, H., Cai, C., & Chen, F. (2017). Automatic classification of traffic incident’s severity using machine learning approaches. IET Intelligent Transport Systems, 11(10), 615-623.
3. Mehar, R., & Agarwal, P. K. (2013). A systematic approach for formulation of a road safety improvement program in India. Procedia-Social and Behavioral Sciences, 104, 1038-1047.
4. Gupta, M., Solanki, V. K., Singh, V. K., & García-Díaz, V. (2018). Data Mining approach of Accident Occurrences Identification with Effective Methodology and Implementation. International Journal of Electrical and Computer Engineering, 8(5), 4033.
5. Caliendo, C., De Guglielmo, M. L., & Russo, I. (2019). Analysis of crash frequency in motorway tunnels based on a correlated random-parameters approach. Tunneling and Underground Space Technology, 85, 243-251.
6. Cai, Q., Saad, M., Abdel-Aty, M., Yuan, J., & Lee, J. (2018). Safety impact of weaving distance on freeway facilities with managed lanes using both microscopic traffic and driving simulations. Transportation Research Record, 2672(39), 130-141.
7. Ifthikar, A., & Hettiarachchi, S. (2018, May). Analysis of historical accident data to determine accident prone locations and cause of accidents. In 2018 8th International Conference on Intelligent Systems, Modelling and Simulation (ISMS) (pp. 11-15). IEEE.
8. Olutayo, V. A., & Eludire, A. A. (2014). Traffic accident analysis using decision trees and neural networks. International Journal of Information Technology and Computer Science, 2, 22-28.
9. Shanthi, S., & Ramani, R. G. (2012, October). Feature relevance analysis and classification of road traffic accident data through data mining techniques. In Proceedings of the World Congress on Engineering and Computer Science (Vol. 1, pp. 24-26).
10. Herceg, L., & Yaman, E. (2019, May). Analysis of Road Accidents using Machine Learning Techniques. In Book of Proceedings (p. 14).
11. Jain, A., Ahuja, G., & Mehrotra, D. (2016, September). Data mining approach to analyse the road accidents in India. In 2016 5th International Conference on Reliability, Infocom Technologies and Optimization (Trends and Future Directions)(ICRITO) (pp. 175-179). IEEE. 15
12. Singh, S. K. (2017). Road traffic accidents in India: issues and challenges. Transportation research procedia, 25, 4708-4719.
13. Kumar, S., & Toshniwal, D. (2016). Analysis of hourly road accident counts using hierarchical clustering and cophenetic correlation coefficient (CPCC). Journal of Big Data, 3(1), 13.
14. Krishnaveni, S., & Hemalatha, M. (2011). A perspective analysis of traffic accident using data mining techniques. International Journal of Computer Applications, 23(7), 40-48.
15. Dwivedi, K., Biswaranjan, K., & Sethi, A. (2014, February). Drowsy driver detection using representation learning. In 2014 IEEE International Advance Computing Conference (IACC) (pp. 995-999). IEEE.
16. Iranitalab, A., & Khattak, A. (2017). Comparison of four statistical and machine learning methods for crash severity prediction. Accident Analysis & Prevention, 108, 27-36.
17. Dogru, N., & Subasi, A. (2018, February). Traffic accident detection using random forest classifier. In 2018 15th Learning and Technology Conference (L&T) (pp. 40-45). IEEE.
18. Júnior, J. F., Carvalho, E., Ferreira, B. V., de Souza, C., Suhara, Y., Pentland, A., & Pessin, G. (2017). Driver behavior profiling: An investigation with different smartphone sensors and machine learning. PLoS one, 12(4), e0174959.
19. Yu, R., & Abdel-Aty, M. (2014). Analyzing crash injury severity for a mountainous freeway incorporating real-time traffic and weather data. Safety science, 63, 50-56.
20. Vasavi, S. (2018). Extracting hidden patterns within road accident data using machine learning techniques. In Information and Communication Technology (pp. 13-22). Springer, Singapore.
21. Kumar, S., & Toshniwal, D. (2016). A data mining approach to characterize road accident locations. Journal of Modern Transportation, 24(1), 62-72.
22. Shanthi, S., & Ramani, R. G. (2012, March). Gender specific classification of road accident patterns through data mining techniques. In IEEE-International Conference On Advances In Engineering, Science And Management (ICAESM-2012) (pp. 359-365). IEEE.
23. Anderson, T. K. (2009). Kernel density estimation and K-means clustering to profile road accident hotspots. Accident Analysis & Prevention, 41(3), 359-364.
24. Parida, M., Jain, S. S., & Kumar, C. N. (2012). Road traffic crash prediction on national highways. Indian Highways, 40(6).
25. Abellán, J., López, G., & De OñA, J. (2013). Analysis of traffic accident severity using decision rules via decision trees. Expert Systems with Applications, 40(15), 6047-6054.
26. Kwon, O. H., Rhee, W., & Yoon, Y. (2015). Application of classification algorithms for analysis of road safety risk factor dependencies. Accident Analysis & Prevention, 75, 1-15.
27. Rizwan, P., Suresh, K., & Babu, M. R. (2016, October). Real-time smart traffic management system for smart cities by using Internet of Things and big data. In 2016 international conference on emerging technological trends (ICETT) (pp. 1-7). IEEE.
28. D’Andrea, E., & Marcelloni, F. (2017). Detection of traffic congestion and incidents from GPS trace analysis. Expert Systems with Applications, 73, 43-56.
29. DOĞRU, N., & SUBAŞI, A. (2015). Comparison of clustering techniques for traffic accident detection. Turkish Journal of Electrical Engineering & Computer Sciences, 23(Sup. 1), 2124-2137.
30. Singh, S. K. (2015). Scenario of urban transport in Indian cities: challenges and the way forward. In Cities and Sustainability (pp. 81-111). Springer, New Delhi.
31. Singh, S. K. (2012). The neglected epidemic: road traffic crashes in India. Metamorphosis, 11(2), 27-49.
32. Abdel-Aty, M., & Haleem, K. (2011). Analyzing angle crashes at unsignalized intersections using machine learning techniques. Accident Analysis & Prevention, 43(1), 461-470.
33. Effati, M., Thill, J. C., & Shabani, S. (2015). Geospatial and machine learning techniques for wicked social science problems: analysis of crash severity on a regional highway corridor. Journal of Geographical Systems, 17(2), 107-135.