This machine learning project focuses on a critical challenge: detecting hazardous asteroids. The primary goal of this endeavor is to construct a robust predictive model with the capability to identify asteroids that pose a threat to Earth. By accurately pinpointing potential hazards, the model aims to enhance decision-making in planetary defense strategies, mitigate potential risks, and facilitate informed actions in the realm of planetary safety.

These are some valuable advantages that are offered by the result of this project:

  • Early Alerts: Timely warnings for potential asteroid threats.
  • Impact Reduction: Minimized risk from hazardous asteroids.
  • Informed Decisions: Data-driven choices for emergency responses.
  • Planetary Safety: Improved defense against celestial hazards.
  • Scientific Insights: Enhanced understanding of asteroid behavior.
  • Public Awareness: Raised awareness and educational opportunities.
  • Global Collaboration: International partnerships for defense efforts.
  • Mission Planning: Optimized space mission targeting.
  • Policy Influence: Guidance for planetary defense policies.
  • Inspiration: Catalyst for innovation and space exploration.

Data Pre-processing

Import used libraries

library(dplyr)
library(caret)
library(inspectdf)
library(Ardian)
library(GGally)
library(e1071)
library(randomForest)

Read the data

The data is from NASA API from http://neo.jpl.nasa.gov/. This API is maintained by SpaceRocks Team: David Greenfield, Arezu Sarvestani, Jason English and Peter Baunach.

asteroids <- read.csv("nasa.csv", stringsAsFactors = T)

Inspect the data

Top 6 rows

asteroids %>% head()

Bottom 6 rows

asteroids %>% tail()

Check duplicated rows

asteroids %>% duplicated() %>% anyNA()
## [1] FALSE

There no duplicated row

Check missing values

asteroids %>% anyNA()
## [1] FALSE

There’s no missing value

Check data structure

asteroids %>% glimpse()
## Rows: 4,687
## Columns: 40
## $ Neo.Reference.ID             <int> 3703080, 3723955, 2446862, 3092506, 35147…
## $ Name                         <int> 3703080, 3723955, 2446862, 3092506, 35147…
## $ Absolute.Magnitude           <dbl> 21.6, 21.3, 20.3, 27.4, 21.6, 19.6, 19.6,…
## $ Est.Dia.in.KM.min.           <dbl> 0.127219878, 0.146067964, 0.231502122, 0.…
## $ Est.Dia.in.KM.max.           <dbl> 0.28447230, 0.32661790, 0.51765448, 0.019…
## $ Est.Dia.in.M.min.            <dbl> 127.219879, 146.067964, 231.502122, 8.801…
## $ Est.Dia.in.M.max.            <dbl> 284.47230, 326.61790, 517.65448, 19.68067…
## $ Est.Dia.in.Miles.min.        <dbl> 0.079050743, 0.090762397, 0.143848705, 0.…
## $ Est.Dia.in.Miles.max.        <dbl> 0.17676284, 0.20295089, 0.32165548, 0.012…
## $ Est.Dia.in.Feet.min.         <dbl> 417.38807, 479.22562, 759.52142, 28.87620…
## $ Est.Dia.in.Feet.max.         <dbl> 933.30809, 1071.58106, 1698.34153, 64.569…
## $ Close.Approach.Date          <fct> 1995-01-01, 1995-01-01, 1995-01-08, 1995-…
## $ Epoch.Date.Close.Approach    <dbl> 788947200000, 788947200000, 789552000000,…
## $ Relative.Velocity.km.per.sec <dbl> 6.115834, 18.113985, 7.590711, 11.173874,…
## $ Relative.Velocity.km.per.hr  <dbl> 22017.004, 65210.346, 27326.560, 40225.94…
## $ Miles.per.hour               <dbl> 13680.510, 40519.173, 16979.662, 24994.84…
## $ Miss.Dist..Astronomical.     <dbl> 0.41948253, 0.38301446, 0.05095602, 0.285…
## $ Miss.Dist..lunar.            <dbl> 163.17871, 148.99263, 19.82189, 110.99039…
## $ Miss.Dist..kilometers.       <dbl> 62753692, 57298148, 7622912, 42683616, 61…
## $ Miss.Dist..miles.            <dbl> 38993336, 35603420, 4736658, 26522368, 37…
## $ Orbiting.Body                <fct> Earth, Earth, Earth, Earth, Earth, Earth,…
## $ Orbit.ID                     <int> 17, 21, 22, 7, 25, 40, 43, 22, 100, 30, 1…
## $ Orbit.Determination.Date     <fct> 2017-04-06 08:36:37, 2017-04-06 08:32:49,…
## $ Orbit.Uncertainity           <int> 5, 3, 0, 6, 1, 1, 1, 0, 0, 0, 6, 4, 6, 0,…
## $ Minimum.Orbit.Intersection   <dbl> 0.025281900, 0.186935000, 0.043057900, 0.…
## $ Jupiter.Tisserand.Invariant  <dbl> 4.634, 5.457, 4.557, 5.093, 5.154, 4.724,…
## $ Epoch.Osculation             <dbl> 2458000, 2458000, 2458000, 2458000, 24580…
## $ Eccentricity                 <dbl> 0.4255491, 0.3516743, 0.3482483, 0.216578…
## $ Semi.Major.Axis              <dbl> 1.4070113, 1.1077760, 1.4588238, 1.255902…
## $ Inclination                  <dbl> 6.025981, 28.412996, 4.237961, 7.905894, …
## $ Asc.Node.Longitude           <dbl> 314.373913, 136.717242, 259.475979, 57.17…
## $ Orbital.Period               <dbl> 609.5998, 425.8693, 643.5802, 514.0821, 4…
## $ Perihelion.Distance          <dbl> 0.8082589, 0.7181996, 0.9507910, 0.983901…
## $ Perihelion.Arg               <dbl> 57.257470, 313.091975, 248.415038, 18.707…
## $ Aphelion.Dist                <dbl> 2.005764, 1.497352, 1.966857, 1.527904, 1…
## $ Perihelion.Time              <dbl> 2458162, 2457795, 2458120, 2457902, 24578…
## $ Mean.Anomaly                 <dbl> 264.837533, 173.741112, 292.893654, 68.74…
## $ Mean.Motion                  <dbl> 0.5905514, 0.8453298, 0.5593708, 0.700277…
## $ Equinox                      <fct> J2000, J2000, J2000, J2000, J2000, J2000,…
## $ Hazardous                    <fct> True, False, True, False, True, False, Fa…

We need to do two things:

  1. Restructure target variable (Hazardous)
  2. Remove uneeded columns

Restructure target variable

asteroids$Hazardous <- ifelse(asteroids$Hazardous == "True", 1, 0) %>% as.factor()

asteroids %>% head(3)

Remove unneeded columns

asteroids <- asteroids %>% 
  select(-Close.Approach.Date,
         -Orbit.Determination.Date,
         -Neo.Reference.ID,
         -Name,
         -Orbit.ID)

asteroids %>% glimpse()
## Rows: 4,687
## Columns: 35
## $ Absolute.Magnitude           <dbl> 21.6, 21.3, 20.3, 27.4, 21.6, 19.6, 19.6,…
## $ Est.Dia.in.KM.min.           <dbl> 0.127219878, 0.146067964, 0.231502122, 0.…
## $ Est.Dia.in.KM.max.           <dbl> 0.28447230, 0.32661790, 0.51765448, 0.019…
## $ Est.Dia.in.M.min.            <dbl> 127.219879, 146.067964, 231.502122, 8.801…
## $ Est.Dia.in.M.max.            <dbl> 284.47230, 326.61790, 517.65448, 19.68067…
## $ Est.Dia.in.Miles.min.        <dbl> 0.079050743, 0.090762397, 0.143848705, 0.…
## $ Est.Dia.in.Miles.max.        <dbl> 0.17676284, 0.20295089, 0.32165548, 0.012…
## $ Est.Dia.in.Feet.min.         <dbl> 417.38807, 479.22562, 759.52142, 28.87620…
## $ Est.Dia.in.Feet.max.         <dbl> 933.30809, 1071.58106, 1698.34153, 64.569…
## $ Epoch.Date.Close.Approach    <dbl> 788947200000, 788947200000, 789552000000,…
## $ Relative.Velocity.km.per.sec <dbl> 6.115834, 18.113985, 7.590711, 11.173874,…
## $ Relative.Velocity.km.per.hr  <dbl> 22017.004, 65210.346, 27326.560, 40225.94…
## $ Miles.per.hour               <dbl> 13680.510, 40519.173, 16979.662, 24994.84…
## $ Miss.Dist..Astronomical.     <dbl> 0.41948253, 0.38301446, 0.05095602, 0.285…
## $ Miss.Dist..lunar.            <dbl> 163.17871, 148.99263, 19.82189, 110.99039…
## $ Miss.Dist..kilometers.       <dbl> 62753692, 57298148, 7622912, 42683616, 61…
## $ Miss.Dist..miles.            <dbl> 38993336, 35603420, 4736658, 26522368, 37…
## $ Orbiting.Body                <fct> Earth, Earth, Earth, Earth, Earth, Earth,…
## $ Orbit.Uncertainity           <int> 5, 3, 0, 6, 1, 1, 1, 0, 0, 0, 6, 4, 6, 0,…
## $ Minimum.Orbit.Intersection   <dbl> 0.025281900, 0.186935000, 0.043057900, 0.…
## $ Jupiter.Tisserand.Invariant  <dbl> 4.634, 5.457, 4.557, 5.093, 5.154, 4.724,…
## $ Epoch.Osculation             <dbl> 2458000, 2458000, 2458000, 2458000, 24580…
## $ Eccentricity                 <dbl> 0.4255491, 0.3516743, 0.3482483, 0.216578…
## $ Semi.Major.Axis              <dbl> 1.4070113, 1.1077760, 1.4588238, 1.255902…
## $ Inclination                  <dbl> 6.025981, 28.412996, 4.237961, 7.905894, …
## $ Asc.Node.Longitude           <dbl> 314.373913, 136.717242, 259.475979, 57.17…
## $ Orbital.Period               <dbl> 609.5998, 425.8693, 643.5802, 514.0821, 4…
## $ Perihelion.Distance          <dbl> 0.8082589, 0.7181996, 0.9507910, 0.983901…
## $ Perihelion.Arg               <dbl> 57.257470, 313.091975, 248.415038, 18.707…
## $ Aphelion.Dist                <dbl> 2.005764, 1.497352, 1.966857, 1.527904, 1…
## $ Perihelion.Time              <dbl> 2458162, 2457795, 2458120, 2457902, 24578…
## $ Mean.Anomaly                 <dbl> 264.837533, 173.741112, 292.893654, 68.74…
## $ Mean.Motion                  <dbl> 0.5905514, 0.8453298, 0.5593708, 0.700277…
## $ Equinox                      <fct> J2000, J2000, J2000, J2000, J2000, J2000,…
## $ Hazardous                    <fct> 1, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0,…

Exploratory Data Analysis

Check target variable proportion

asteroids$Hazardous %>% table() %>% barplot()

The target proportion is imbalanced. Let’s upsample!

Upsample the data

up_asteroids <- upSample(x = asteroids %>% select(-Hazardous),
                         y = asteroids$Hazardous,
                         yname = "Hazardous")

up_asteroids$Hazardous %>% table() %>% barplot()

Cool. It’s balanced now!

Inspect categorical columns distributions

up_asteroids %>% inspect_cat() %>% show_plot()

The column Equinox and Orbiting.Body has 0 variance. Let’s remove them!

Remove zero variance columns

up_asteroids <- up_asteroids %>% select(-Equinox,
                                        -Orbiting.Body)

up_asteroids %>% head(3)

Inspect numerical columns distributions

up_asteroids %>% select_if(is.numeric) %>% summary()
##  Absolute.Magnitude Est.Dia.in.KM.min.  Est.Dia.in.KM.max. Est.Dia.in.M.min.  
##  Min.   :11.16      Min.   : 0.001011   Min.   : 0.00226   Min.   :    1.011  
##  1st Qu.:19.60      1st Qu.: 0.076658   1st Qu.: 0.17141   1st Qu.:   76.658  
##  Median :21.00      Median : 0.167708   Median : 0.37501   Median :  167.708  
##  Mean   :21.40      Mean   : 0.249889   Mean   : 0.55877   Mean   :  249.889  
##  3rd Qu.:22.70      3rd Qu.: 0.319562   3rd Qu.: 0.71456   3rd Qu.:  319.562  
##  Max.   :32.10      Max.   :15.579552   Max.   :34.83694   Max.   :15579.552  
##  Est.Dia.in.M.max.  Est.Dia.in.Miles.min. Est.Dia.in.Miles.max.
##  Min.   :    2.26   Min.   :0.000628      Min.   : 0.001404    
##  1st Qu.:  171.41   1st Qu.:0.047633      1st Qu.: 0.106510    
##  Median :  375.01   Median :0.104209      Median : 0.233019    
##  Mean   :  558.77   Mean   :0.155274      Mean   : 0.347203    
##  3rd Qu.:  714.56   3rd Qu.:0.198566      3rd Qu.: 0.444008    
##  Max.   :34836.94   Max.   :9.680682      Max.   :21.646663    
##  Est.Dia.in.Feet.min. Est.Dia.in.Feet.max. Epoch.Date.Close.Approach
##  Min.   :    3.32     Min.   :     7.41    Min.   :7.889e+11        
##  1st Qu.:  251.50     1st Qu.:   562.37    1st Qu.:9.905e+11        
##  Median :  550.22     Median :  1230.34    Median :1.182e+12        
##  Mean   :  819.85     Mean   :  1833.23    Mean   :1.164e+12        
##  3rd Qu.: 1048.43     3rd Qu.:  2344.36    3rd Qu.:1.344e+12        
##  Max.   :51114.02     Max.   :114294.42    Max.   :1.473e+12        
##  Relative.Velocity.km.per.sec Relative.Velocity.km.per.hr Miles.per.hour   
##  Min.   : 0.3355              Min.   :  1208              Min.   :  750.5  
##  1st Qu.: 9.6670              1st Qu.: 34801              1st Qu.:21624.2  
##  Median :14.2036              Median : 51133              Median :31772.0  
##  Mean   :15.2773              Mean   : 54998              Mean   :34173.7  
##  3rd Qu.:19.5768              3rd Qu.: 70476              3rd Qu.:43791.3  
##  Max.   :44.6337              Max.   :160681              Max.   :99841.2  
##  Miss.Dist..Astronomical. Miss.Dist..lunar.   Miss.Dist..kilometers.
##  Min.   :0.0001779        Min.   :  0.06919   Min.   :   26610      
##  1st Qu.:0.1391497        1st Qu.: 54.12924   1st Qu.:20816502      
##  Median :0.2732463        Median :106.29281   Median :40877064      
##  Mean   :0.2626487        Mean   :102.17036   Mean   :39291694      
##  3rd Qu.:0.3901572        3rd Qu.:151.77116   3rd Qu.:58366688      
##  Max.   :0.4998841        Max.   :194.45491   Max.   :74781600      
##  Miss.Dist..miles.  Orbit.Uncertainity Minimum.Orbit.Intersection
##  Min.   :   16535   Min.   :0.000      Min.   :0.0000021         
##  1st Qu.:12934775   1st Qu.:0.000      1st Qu.:0.0127533         
##  Median :25399830   Median :1.000      Median :0.0296974         
##  Mean   :24414727   Mean   :2.578      Mean   :0.0582420         
##  3rd Qu.:36267380   3rd Qu.:6.000      3rd Qu.:0.0642087         
##  Max.   :46467132   Max.   :9.000      Max.   :0.4778910         
##  Jupiter.Tisserand.Invariant Epoch.Osculation   Eccentricity     
##  Min.   :2.196               Min.   :2450164   Min.   :0.007522  
##  1st Qu.:4.069               1st Qu.:2458000   1st Qu.:0.269198  
##  Median :5.085               Median :2458000   Median :0.408658  
##  Mean   :5.053               Mean   :2457756   Mean   :0.413052  
##  3rd Qu.:5.990               3rd Qu.:2458000   3rd Qu.:0.546893  
##  Max.   :9.025               Max.   :2458020   Max.   :0.960261  
##  Semi.Major.Axis   Inclination       Asc.Node.Longitude Orbital.Period  
##  Min.   :0.6159   Min.   : 0.01451   Min.   :  0.0019   Min.   : 176.6  
##  1st Qu.:1.0061   1st Qu.: 4.85172   1st Qu.: 84.3511   1st Qu.: 368.6  
##  Median :1.2332   Median : 9.97538   Median :174.1097   Median : 500.2  
##  Mean   :1.3925   Mean   :13.46782   Mean   :174.3131   Mean   : 629.7  
##  3rd Qu.:1.6554   3rd Qu.:19.80588   3rd Qu.:259.4328   3rd Qu.: 777.9  
##  Max.   :5.0720   Max.   :75.40667   Max.   :359.9059   Max.   :4172.2  
##  Perihelion.Distance Perihelion.Arg     Aphelion.Dist    Perihelion.Time  
##  Min.   :0.08074     Min.   :  0.0069   Min.   :0.8038   Min.   :2450100  
##  1st Qu.:0.58926     1st Qu.: 97.4822   1st Qu.:1.3002   1st Qu.:2457826  
##  Median :0.78809     Median :193.9037   Median :1.6712   Median :2457983  
##  Mean   :0.76705     Mean   :184.6180   Mean   :2.0180   Mean   :2457758  
##  3rd Qu.:0.95089     3rd Qu.:268.1775   3rd Qu.:2.4768   3rd Qu.:2458108  
##  Max.   :1.29983     Max.   :359.9931   Max.   :8.9839   Max.   :2458839  
##   Mean.Anomaly       Mean.Motion     
##  Min.   :  0.0032   Min.   :0.08628  
##  1st Qu.: 96.2437   1st Qu.:0.46277  
##  Median :191.2148   Median :0.71971  
##  Mean   :187.2033   Mean   :0.74210  
##  3rd Qu.:281.9666   3rd Qu.:0.97666  
##  Max.   :359.9180   Max.   :2.03900

All numerical columns are distributed normally. Let’s move on to cross validation!

Cross Validation

Set training indices

Because we don’t have a lot of records, we will set the training proportion to 85%

set.seed(1)
indices <- createDataPartition(y = up_asteroids$Hazardous,
                               p = 0.85,
                               list = F)

Split train & test

train_data <- up_asteroids[indices, ]
test_data <- up_asteroids[-indices, ]

X_train <- train_data %>% select(-Hazardous)
X_test <- test_data %>% select(-Hazardous)

y_train <- train_data$Hazardous
y_test <- test_data$Hazardous

Model Fitting, Evaluation, & Selection

Naive Bayes Algorithm

model_nb <- naiveBayes(x = X_train,
                       y = y_train,
                       laplace = 1)

Naive Bayes Model Evaluation

pred_nb <- predict(model_nb, X_test)

confusionMatrix(pred_nb, y_test)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction   0   1
##          0 549  35
##          1  40 554
##                                           
##                Accuracy : 0.9363          
##                  95% CI : (0.9208, 0.9496)
##     No Information Rate : 0.5             
##     P-Value [Acc > NIR] : <2e-16          
##                                           
##                   Kappa : 0.8727          
##                                           
##  Mcnemar's Test P-Value : 0.6442          
##                                           
##             Sensitivity : 0.9321          
##             Specificity : 0.9406          
##          Pos Pred Value : 0.9401          
##          Neg Pred Value : 0.9327          
##              Prevalence : 0.5000          
##          Detection Rate : 0.4660          
##    Detection Prevalence : 0.4958          
##       Balanced Accuracy : 0.9363          
##                                           
##        'Positive' Class : 0               
##

The focused metric for our evaluation is Accuracy. Why? Because in the realm of hazardous asteroid detections, it’s crucial to ensure a reliable overall performance in identifying both hazardous and non-hazardous instances.

The Naive Bayes model has demonstrated an overall Accuracy of 92%, accompanied by a Sensitivity of 93% and Specificity of 92%. The high Sensitivity indicates the model’s effectiveness in correctly identifying hazardous asteroids, a critical capability for early detection and accurate assessment of potential threats. Furthermore, the impressive Specificity underscores the model’s ability to accurately classify non-hazardous asteroids. These results collectively contribute to our ability to proactively mitigate the risk posed by hazardous celestial bodies and enhance our planetary safety measures.

NOTE: Due to the unpredictable nature of R Markdown rendering, the generated scores might differ when the document is converted to HTML

Logistic Regression Algorithm

model_lgr <- glm(Hazardous ~ ., "binomial", train_data)

Logistic Regression Evaluation

pred_lgr <-  ifelse(predict(model_lgr, X_test) > 0.5, 1, 0) %>% as.factor()

confusionMatrix(pred_lgr, y_test)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction   0   1
##          0 558  37
##          1  31 552
##                                           
##                Accuracy : 0.9423          
##                  95% CI : (0.9274, 0.9549)
##     No Information Rate : 0.5             
##     P-Value [Acc > NIR] : <2e-16          
##                                           
##                   Kappa : 0.8846          
##                                           
##  Mcnemar's Test P-Value : 0.5443          
##                                           
##             Sensitivity : 0.9474          
##             Specificity : 0.9372          
##          Pos Pred Value : 0.9378          
##          Neg Pred Value : 0.9468          
##              Prevalence : 0.5000          
##          Detection Rate : 0.4737          
##    Detection Prevalence : 0.5051          
##       Balanced Accuracy : 0.9423          
##                                           
##        'Positive' Class : 0               
##

The logistic regression model exhibited an impressive Accuracy of 93%, alongside a Sensitivity of 95% and Specificity of 92%. The high Sensitivity demonstrates the model’s effectiveness in correctly flagging hazardous asteroids, which is of paramount importance for early detection and mitigation efforts, contributing to enhanced planetary safety.

Random Forest Algorithm

ctrl <- trainControl(method = "repeatedcv",
                     number = 3,
                     repeats = 5)

model_rf <- train(Hazardous ~ .,
                  data = train_data,
                  method = "rf",
                  trConrol = ctrl,
                  ntree = 101)

Random Forest Model Evaluation

pred_rf <- predict(model_rf, X_test)

confusionMatrix(pred_rf, y_test)
## Confusion Matrix and Statistics
## 
##           Reference
## Prediction   0   1
##          0 589   0
##          1   0 589
##                                      
##                Accuracy : 1          
##                  95% CI : (0.9969, 1)
##     No Information Rate : 0.5        
##     P-Value [Acc > NIR] : < 2.2e-16  
##                                      
##                   Kappa : 1          
##                                      
##  Mcnemar's Test P-Value : NA         
##                                      
##             Sensitivity : 1.0        
##             Specificity : 1.0        
##          Pos Pred Value : 1.0        
##          Neg Pred Value : 1.0        
##              Prevalence : 0.5        
##          Detection Rate : 0.5        
##    Detection Prevalence : 0.5        
##       Balanced Accuracy : 1.0        
##                                      
##        'Positive' Class : 0          
##

The random forest model demonstrated remarkable performance, achieving an exceptional Accuracy of 100%, making it the winner model of this project!. Furthermore, the model achieved a Sensitivity of 100% and Specificity of 100%. This outstanding performance reflects the model’s ability to flawlessly distinguish hazardous asteroids from non-hazardous ones. Such high sensitivity and specificity are instrumental in identifying and classifying objects accurately, underscoring the model’s reliability in predicting hazardous asteroids. This level of precision is of paramount importance for effective early detection and risk assessment, contributing to enhanced planetary safety and preparedness.

AUC of ROC

My way of explaining the AUC of ROC score is that it reflects the level of certainty our model has in its predictions. A high AUC of ROC score indicates that the model is very confident and sure of its predictions. As people who use the model, we want the model to be highly confident in its predictions since we rely on it for making decisions. We don’t want a model that isn’t sure or confident about its predictions. This is why the AUC or ROC score is a crucial measure to determine if the model is prepared for practical use or not.

pred_rf_raw <- predict(model_rf, X_test, type = "prob")

plotROC(pred_rf_raw[, 2], y_test) # This function is from my personal package

Magnificent. Why? Because the closer the AUC to 1 the more confident the model is at detecting which asteroids are hazardous and which are not

Conclusion

In conclusion, the final model developed for hazardous asteroid detection, which is the random forest model, stands as an exceptional achievement in predictive modeling. With an extraordinary Accuracy of 100% and Sensitivity of 100%, the model showcases unparalleled precision in identifying hazardous asteroids. This model also demonstrates perfect Specificity of 100%, further highlighting its ability to accurately differentiate non-hazardous asteroids.

Moreover, the AUC of ROC score of 1.0, representing a perfect score, serves as a resounding validation of the model’s readiness for practical deployment. This impeccable AUC score underscores the model’s absolute confidence in distinguishing between hazardous and non-hazardous asteroids. The model’s remarkable performance across all evaluation metrics solidifies its position as a reliable and robust tool for effectively detecting potential asteroid threats.

The flawless performance achieved by the random forest model, marked by its perfect accuracy, sensitivity, and specificity, attests to its exceptional competence in the realm of hazardous asteroid detection. With its unprecedented performance metrics, the model stands as a testament to its preparedness for real-world application, offering invaluable contributions to planetary defense and safety.