Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization
The present analysis of state of the art portrays that actual time series or spectrum backscattered data from a point on the sea bottom are rarely used as features for machine learning models. The paper deals with the artificial intelligence techniques used to examine CHIRP-recorded data. The data w...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-07-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/14/8131 |
| _version_ | 1797590427788902400 |
|---|---|
| author | Uri Kushnir Vladimir Frid |
| author_facet | Uri Kushnir Vladimir Frid |
| author_sort | Uri Kushnir |
| collection | DOAJ |
| description | The present analysis of state of the art portrays that actual time series or spectrum backscattered data from a point on the sea bottom are rarely used as features for machine learning models. The paper deals with the artificial intelligence techniques used to examine CHIRP-recorded data. The data were collected using a CHIRP sub-bottom profiler to study two sand bottom sites and two sandstone bottom sites in the offshore zone of Ashqelon City (Southern Israel). The first reflection time series and spectra of all the traces from the four sites generated two training and two test sets. Two logistic regression models were trained using the training sets and evaluated for accuracy using the test sets. The examination results indicate that types of sea bottom can be quantitatively characterized by applying logistic regression models to either the backscatter time series of a frequency-modulated signal or the spectrum of that backscatter. The examination accuracy reached 90% for the time series and 94% for the spectra. The application of spectral data as features for more advanced machine learning algorithms and the advantages of their combination with other types of data have great potential for future research and the enhancement of remote marine soil classification. |
| first_indexed | 2024-03-11T01:20:19Z |
| format | Article |
| id | doaj.art-1f0867b58c07400f9d29e52f7e6effa5 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T01:20:19Z |
| publishDate | 2023-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-1f0867b58c07400f9d29e52f7e6effa52023-11-18T18:08:42ZengMDPI AGApplied Sciences2076-34172023-07-011314813110.3390/app13148131Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil CategorizationUri Kushnir0Vladimir Frid1Civil Engineering Department, Sami Shamoon College of Engineering, Ashdod 77245, IsraelCivil Engineering Department, Sami Shamoon College of Engineering, Ashdod 77245, IsraelThe present analysis of state of the art portrays that actual time series or spectrum backscattered data from a point on the sea bottom are rarely used as features for machine learning models. The paper deals with the artificial intelligence techniques used to examine CHIRP-recorded data. The data were collected using a CHIRP sub-bottom profiler to study two sand bottom sites and two sandstone bottom sites in the offshore zone of Ashqelon City (Southern Israel). The first reflection time series and spectra of all the traces from the four sites generated two training and two test sets. Two logistic regression models were trained using the training sets and evaluated for accuracy using the test sets. The examination results indicate that types of sea bottom can be quantitatively characterized by applying logistic regression models to either the backscatter time series of a frequency-modulated signal or the spectrum of that backscatter. The examination accuracy reached 90% for the time series and 94% for the spectra. The application of spectral data as features for more advanced machine learning algorithms and the advantages of their combination with other types of data have great potential for future research and the enhancement of remote marine soil classification.https://www.mdpi.com/2076-3417/13/14/8131marine surveyacoustic reflectionspectral analysissediments identification |
| spellingShingle | Uri Kushnir Vladimir Frid Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization Applied Sciences marine survey acoustic reflection spectral analysis sediments identification |
| title | Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization |
| title_full | Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization |
| title_fullStr | Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization |
| title_full_unstemmed | Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization |
| title_short | Spectrum-Based Logistic Regression Modeling for the Sea Bottom Soil Categorization |
| title_sort | spectrum based logistic regression modeling for the sea bottom soil categorization |
| topic | marine survey acoustic reflection spectral analysis sediments identification |
| url | https://www.mdpi.com/2076-3417/13/14/8131 |
| work_keys_str_mv | AT urikushnir spectrumbasedlogisticregressionmodelingfortheseabottomsoilcategorization AT vladimirfrid spectrumbasedlogisticregressionmodelingfortheseabottomsoilcategorization |