Assignment No 2 AI techniques forNatural Disasters Prediction By: Your NameTable of Contents1- Natural disaster 32- Natural disaster prediction 33- AI techniques for Earthquake predictions 33.1- Feed forward neural network 43.2- Particle swarm optimization 43.
3- Genetic Algorithm 43.4- Clustering 54- AI techniques for Water stormpredictions 54.1- Nonlinear ensemble prediction 54.2- Back propagation neural network 65- Degree of success in natural disasterprediction by AI 76- Summary 87- References 9 1- Naturaldisaster It is a natural event such as flood, earthquake or hurricanethat cause damage or loss of life. It effects both living and non-living.
2- NaturalDisaster predictionNatural disasters are inevitable in our world. Naturaldisasters are of different types so it is difficult to predict each and everyone. Meteorologists can track a hurricane with precision, butseismologists cannot predictexactly when and where an earthquake will occur. Prediction of disasters require extensive research andfunding. To predict a natural disaster we have to collect extensive past data,record live data and generate patterns on previous data. By comparing past andlive data scientist predicts the future events to some extent. Trends arecalculated and used to predict earthquakes, tsunamis and volcanic eruptions.
We can also predict natural disasters by constantsurveillance. Using offshore cameras in hurricane prone areas ensures thatstrong winds and waves can be recognized, that will help in tsunami predictions.Monitoring ocean currents, weather predictions can be predicted in advance,warning nearby areas in advance under the risk of hurricanes and tornados. Butthese short term warnings are only effective when relief programs are plannedand effectively carried out. But this method is very costly and inefficient.For cost effectiveness and timely information of naturaldisaster, predicting it in advance is the only solution. However it is notalways reliable, because disasters unexpected and do not always follows trends.But it will save much time and resource than constant surveillance.
3- AItechniques for earthquake predictionsNatural disasters like earthquakes are caused due to the propagatingseismic waves underneath the surface of earth. Seismometers are installed ondifferent geographical positions to record vertical motion of surface waves. Groundmotion types are divergence, convergence which results in transforming plateboundaries. Major earthquakes are caused by divergence, convergence andtransformation of plate boundaries commonly known as faults. The origin where earthquake takes place isorigin point. Total sum of waves are calculated and time series data iscollected for further processing.There are four different aspects of this time series datawith respect to geophysical analysis can be considered for experimentation.
1. Analyze the earthquake data recording indifferent time points independent of common source gather or common receivergather recordings.2. Analyze the earthquake data set in fixed orvariable length time intervals to predict different hidden patterns3.
Gathering layers data, like layer between Euro-Asianand Indian plate etc, in time points to better analyze and study the seismicpatterns of layer with respect to time4. Gather and analyze the earth lithosphere layerdata with respect to time intervals Such identified characteristics of earthquake can be easilyscaled down using some activation function. Figure 1: Illustration of criteria for fitness function3.1 – Feed Forward Neural NetworkIt is used with sigmoid function. FFN is used on Seismic Electric signals, predicted magnitude andpre-determined future seismic events. Prediction of structural responses for astructure has 80.55% accuracy.
Prediction has 71%efficiency. It is ableto predict both long and short term shocks. Outputs of different layers are notfeedback. 3.2 – Particle Swarm OptimizationPSO is used for building prior knowledge system. It is used for selection of input valuesfor the BPN (Back Propagation NeuralNetwork) based network. It can determine earthquake local earthquakelocation. Works on the principles of Swarms of particlessearching for optimal solution in the defined search space.
Converge to thesolution more efficiently then general BPN. 3.3 – Genetic Algorithm Rock mass stability is estimated for planningpurpose. Structural formation has been studied using GA. Lower the data uncertainty.
Used for buildingsettlement forecast after main shocks. Used in combination with support vectormachines for earthquake data set. GA can work with improper or incomplete seismicdata. It is found highly efficient in prediction for future earthquakes.
Commonly used in research with different alterations. 3.4 – ClusteringSpatial clustering is used versus temporalclustering for earthquake data sets. Spatial clustering has been identified indata set while building earthquake forecast model using differentialprobability. Set of clusters is developed from huge set ofunsupervised data. This makes the overall scenario to be divided into manysub-scenarios. Used in MSc algorithm with different aspect. Until recently,Artificial Intelligence based techniques were widely used for earthquake timeseries prediction.
, the results of traditional approaches of probabilityestimation should be enhanced by using the particle swarm optimization andgenetic algorithms based approaches. PSO and GA are capable to find actualfault intensity in any particular region. This work is an attempt to coverdifferent strategies related to AI for earthquake prediction and crosschecktheir reliability. . 4- AItechniques for Water Storms PredictionWater storms occur due to intenseunsustainable winds in oceans. Hurricanes, cyclones and typhoons are all water stormsbut their name is different due to the geographical location of storm.
There are different artificialintelligence techniques to predict storms. Some are given below. 4.1 – Nonlinear AI ensemblepredictionA new nonlinear artificialintelligence ensemble prediction (NAIEP) model has been developed forpredicting typhoon intensity based on multiple neural networks with the sameexpected output and using an evolutionary genetic algorithm (GA). Ensemble numerical prediction(ENP) model, whether created with different physical process parameterizationschemes or with different initial conditions from a Monte Carlo approach,formally consists of many different ensemble members.
By optimizing the networkstructure and the connection weight of ANNs, genetic evolution is able tocreate a number of different neural network individuals.Ensemble prediction of NWP ismotivated by the fact that NWP forecasts are sensitive both to smalluncertainties in the initial conditions and model errors, so it is hard tofurther improve the accuracy of single model deterministic predictions.To construct an NAIEP model, anumber of individual neural networks are first created and then integrated tobuild an ensemble prediction model.A GA is used to construct themembers of the ensemble, and a three-layer back-propagation (BP) network isused as the basic model for the neural networks, the major computational stepsare summarized below:1. Randomlygenerate the connection weights and thresholds from input layer to hidden layerand from hidden layer to output layer, and set the global convergence error, ?,of the model.2. Performsupervised learning training of the network with learning matrix samples,calculate the error between the real input and expected output of the network,and tune the connection weight coefficients from input layer to hidden layerand from hidden layer to output layer using the learning algorithm of theerror-inverse propagation of the BP network.3.
If thecalculated output error of the model is greater than ?, return to step 2;otherwise, end the training and compute the prediction value using theconnection weights, thresholds of the network, and predictors of the predictionsamples.The meteorological ensemblemodeling approach of GNN opens up a vast range of possibilities for operationalweather prediction. 4.2 – Back propagation Neural NetworkLike human neural network inartificial neural network has 3 layers; perceptron, dendrites and axon. In NN,each input is multiplied by its weight of its connection of neuron.
Connectiondetermine which input has to be forward and then it sums up all the inputs.Then it is passed through the hidden layer to calculate its results. After itpasses the result to output layer. In back propagation NN, there is only oneinput layer, one output layer and a hidden layer. It is easy to calcite theresults.To compute the prediction ofstorm or any other disaster information is collected and then it is feed to theneural network. First of all data is normalized then it is feed to input layer.From input layer the data is transferred to hidden layer.
There we do ourcalculations by applying sigmoid function. From hidden layer calculations arecollected and summed up, this sum is input to output layer. NN with back propagation is aself-driving system which collects data then train itself for differentconditions and scenarios and produce results.NN with back propagation andother NN’s are not more than pattern recognition techniques.
They are just someshort term predictive skills not to replace metrologies. But it can help inunderstanding metrological problems and can solve many complicated patternsthat are difficult to solve by humans and simple programs. 5- Degreeof success in natural disaster prediction by AI For some people weather forecastsare just for surety of good day ahead. But for some people it is everything.
Their bread and butter depends upon it. By applying artificial intelligenceknowledge we have been able to transforms life of many people and giving them anew chance. Different companies andgovernments are collecting data of winds, water and soil from satellites anddifferent devices installed on earth. By applying artificial intelligence witha physical understanding of environment can significantly improve theprediction skill for multiple types of high-impact weather. High-impact weatherincludes events like severe thunderstorms, tornadoes, and hurricanes.One example the paper highlightedis that machine learning can provide more accurate hail forecasting. Hail cause billions of dollars damages every year. Even a modestimprovement in hail warning could produce significant savings by gettingindividuals move their cars and themselves to safety.
Provide these types ofwarnings for car insurance companies is one way IBM is commercializing theirweather predictions. IBM, Panasonic are workingheavily on their weather forecasting systems and improving them day by day byapplying new developed artificial intelligence and computational intelligencetechniques. Better weather forecasting allows airlines to adjust their routesto reduce fuel use, improve safety and increase on time arrives.Better weather predictions hasdirect effect on different fields of life. It directly effects the agriculture,90% of crops are destroyed by weather conditions.
It can be controlled byproper weather conditions. If damage is inevitable then we can save our moneyand time by not planting that kind of crops that are not suitable for that kindof weather condition. Transportation is improved by weatherforecasting. By directing routes and stopping the flights in near storm savemany lives and money than ever. The company Safety Line is using Panasonic’s weather forecastingto optimize the climb profile for commercial aircraft. They claim their systemcan reduce fuel consumption by up to 10 percent during ascent.
To put that inperspective, US airline carriers spent $24.6 billion on fuel last year.Better weather forecasting saveslives and helps speed up rebuilding efforts.
IBM has started combining theirweather forecasting tools with information about utilities’ distributionnetworks and data about local ground cover for severe storms. Using machinelearning, they are predicting likely outages. IBM claims they deliver damage predictions thatare 70 to 80 percent accurate 72 hours before the storm is expected.The potential source ofweather-related data will continue to grow dramatically and the new advances inmachine learning are making it possible for government agencies and companiesto make better use of all this data. Weather forecasting can never be trulyperfect, but AI will allow the practice to continue to improve in its accuracyand in its resolution.The more refined and localizedour weather information gets, the easier it will be to find distinct patternsand connections. Even small improvements in weather forecasting will givecompanies new useful pieces of data by finding new correlations and givingcompanies more lead time to take advantage of them. 6- SummaryNatural disasters are inevitableand unpredicted in nature.
Nobody can exactly predict that what will happennext. But by the passage of time human being is able to extract informationfrom past events and made patterns from that information. In past those patternswere not so clear and difficult to compute. Modern day technology helps tocollect data and draws results from that data effectively. Artificial intelligence plays akey role in pattern recognition and analysis of past events to predict futureevents. Different techniques like Neural network (NN), genetic algorithm (GN),particle swarm optimization (PSO), clustering and many more helps us to find patternsand prediction.
These algorithms alone do not generate good results. But by mergingtwo algorithms gives us better results that give us better understanding of occurringand future events. By using artificial intelligencetechniques success rate of prediction is about 60-70%. Although it is not so accurateone but it helps to save resources and lives. By using weather prediction cropsare not destroying any more than before.
Air traffic is controlled in a goodmanner and they are informed before any bad can happen to them. In short, artificialintelligence made a good impact on the life of people by giving them usefulpiece of information in advance. 7- ReferencesAberson, S. D., and C.
R. Sampson, 2003:On the predictability of tropical cyclone tracksin the northwest Pacific basin. Mon. Wea.
Rev., 131, 1491–1497. Chen, G-L., X-F. Wang, and Z-Q. Zhuang, 1996: Genetic Algorithm and Application. (in Chinese). Beijing Communication Press, 178 pp.
Allen J.F. Time and time again: the many ways to represent time.
InternationalJournal of Intelligent Systems. 1991;6(4):341–355.doi: 10.1002/int.4550060403BreimanL. Random forests. Machine Learning. 2001;45(1):5–32.
doi: 10.1023/A:1010933404324.Davies-JonesR. P. Tornado dynamics.
In: Kessler E., editor. Thunderstormmorphology and dynamics.
Norman: University of OklahomaPress; 1986. pp. 197–236.
U.S. Dept. of Commerce/NO A A: National Hurricane OperationPlan: Federal Coordination For Metrology Services and Supporting Research,Washington D.C. 1992. D. E.
Rumelhart, J. L. McClelland, “Exploration in theMicrostructure of Cognition” in Parallel Distributed Processing, The MITPress, vol. 1, 1986S. Y.
Pakkala, F. C. Lin, Proceedings of the SPIEConference,1992-April-2124.DouillyR., Haase J.
S., Ellsworth W. L., Bouin M.
P., Calais E., Symithe S. J.
,Armbruster J. G., De Lépinay B.
M., Deschamps A. and Mildor S. L.
, CrustalStructure and Fault Geometry of the 2010 Haiti Earthquake from Temporary SeismometerDeployments, Bulletin of the Seismological Society of America,103(4),2305-2325 (2013)Klein E. M., Geochemistry of theIgneous Oceanic Crust, Treatise on Geochemistry, 3, 433-463 (2003)Holtzman B. K. and Kendall J.,Organized melt, seismic anisotropy, and plate boundary lubrication, Geochemistry,Geophysics, Geosystems, 11(12), 1-29 (2010)