Figure 4-1: Profile Management function
4.2.2. Function to update general profile information
When the appraiser wants to create a new property appraisal file. The appraiser selects Add new file. First, the appraiser needs to enter general information about the file.
Figure 4-2: Update general profile information.
4.2.3. Update legal information of the file
After updating general information about the file, the specialist needs to fully declare legal information related to the property in the appraisal file.
Figure 4-3: Updating legal information of assets in appraisal records
4.2.4. Update information on collateral
The function requires the specialist to update the information of the collateral assets that need to be appraised in the file.
Figure 4-4: Update information on collateral - Real estate
4.2.5. Update the conclusion information of the appraisal report
After completing the appraisal, the specialist will update the information concluding the value of the appraised assets in the file.
Figure 4-5: Conclusion of the asset appraisal report
4.2.6. Price explanation
The function supports users in making price decisions for assets. With the elements that users enter to search. The system uses the KNN algorithm to determine 2 data sets: 3 comparable assets with the most similar elements to the asset to be appraised and all assets with the most similar elements to the information entered by the user. Visual display in the form of charts and tables helps users easily compare and track. In addition, the system automatically builds a multivariate linear regression model with the important elements of the asset, on the data set just found by the KNN algorithm.
Figure 4-6: Price description
4.2.7. Comparison of experimental results on Web assessment and Weka software
Input: The test is performed on 10 property information in the dataset in Appendix 1. With the extracted attributes: Width, Length, Lane frontage, Land orientation, Location (Longitude, Latitude).
Compare the two methods:
- Method 1 with result column 1: Experimental results on the Collateral Valuation website, using a combination of the K-nearest neighbor method and multivariate linear regression.
- Method 1 with result column 2: Experimental results on Weka software, predicted by multivariate linear regression method.
Output:
- Multivariate linear regression equation performed on the website for collateral appraisal:
Property value = -4546 + 576*Width + 369*Length - 167*Alley Frontage -
9*Land Direction
Variance = 0.87
- Multivariate linear regression equation using multivariate linear regression method in weka software:
Property value = -4786+ 0*Width + 265*Length +187*Alley Frontage
+5756.0096 * Direction=Southeast,West
Variance = 0.53
- The test results are as follows:
Table 4-1: Test results of two methods
Resources
product
Location | Afternoon horizontal | Afternoon long | Face Alley | Direction land | Price reality | Result 1 | Result 2 | |
1 | Hiep Binh Ward Chanh, Thu Duc, Ho Chi Minh | 6.0 | 12.7 | 10.0 | Southeast | 2579.0 | 1851.0 | 6175 |
2 | Street 27, Hiep Binh Chanh Ward, Thu Duc, Ho Chi Minh Bright | 30.0 | 26.0 | 12.0 | Southeast | 23340.0 | 20262.0 | 13704 |
Maybe you are interested!
-
Issue Two: Appraisal of Collateral -
Resolving the Contradiction Between the Relatively Low Educational Level of Many Prisoners and the Requirements for Acquiring Legal Information and Knowledge -
Right to access information under Vietnamese law and compatibility with international legal standards - 10 -
Factors That Hinder the Process of Handling Collateral Assets to Recover Bad Debts at Maritime Commercial Joint Stock Bank, Hanoi Branch -
Applying Business Intelligence in the problem of bank collateral appraisal - 12
3
Ngo Chi Quoc Street, Binh Chieu Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 6.0 | 11.7 | 12.0 | Southeast | 1610.0 | 1148.0 | 6129 | |
4 | Ngo Chi Quoc Street, Binh Chieu Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 5.0 | 12.4 | 12.0 | Southeast | 1426 | 830 | 6500 |
5 | Ngo Chi Quoc Street, Binh Chieu Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 12.0 | 14.8 | 12.0 | Southeast | 4200 | 5751 | 7136 |
6 | Provincial Road 43, Binh Chieu Ward, Thu Duc, Ho Chi Minh Bright | 5.0 | 10.0 | 5.0 | Southeast | 1400 | 1113 | 4555 |
7 | National Highway 13, Hiep Binh Phuoc Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 4.6 | 11.5 | 7.0 | North | 1470 | 1111 | -429 |
8 | National Highway 13, Hiep Binh Phuoc Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 4.2 | 12.6 | 6.0 | Northwest | 1456 | 1463 | -325 |
9 | National Highway 13, Hiep Binh Phuoc Ward, Thu Duc, Ho Chi Minh City Ho Chi Minh | 6.0 | 7.0 | 7.0 | Southeast | 1500 | 2470 | 4134 |
10 | Linh Chieu Ward, Thu Duc District, Ho Chi Minh City Ho Chi Minh | 5 | 10 | 8 | Male | 820 | 616 | -640 |
Comment:
- The predicted asset value when using the multivariate linear regression method on weka software (Result column 2) has a much larger difference from the actual value than the result obtained from the Appraisal Web (Result column 1).
- The variance value of method 1 is larger than that of method 2. This shows that the predicted value of method 1 is more reliable.
=> Summary: The combination of two algorithms K nearest neighbors and multivariate linear regression gives more accurate prediction results than using only linear regression on the same data set.