Customer Behavior Classification Using Deep Stacked Autoencoder with Dragonfly Optimization
Uloženo v:
| Název: | Customer Behavior Classification Using Deep Stacked Autoencoder with Dragonfly Optimization |
|---|---|
| Autoři: | Gajendran K S, Arunkumar Thangavelu |
| Zdroj: | Journal of Machine and Computing. :2230-2240 |
| Informace o vydavateli: | Anapub Publications, 2025. |
| Rok vydání: | 2025 |
| Popis: | Customer Relationship Management (CRM) plays a major role in analyzing customer behavior and the opinions of an organization or enterprise. Data mining methods are widely uutilize to analyze customer data to increase business and revenue. Data mining refers to the extraction of essential and useful information from customer feedback and activities on websites through mining technologies. However, extracting essential information from customer behavior is quite challenging as it requires a detailed analysis of customer desires, requirements, buying patterns, etc., all the information in the e-commerce market is essential for an enterprise as it will support knowing the customer behavior. Deep learning algorithms based on customer behavior classification models are evolved in recent times. However, the performance can be improved if the network parameters are optimized through optimization algorithms. Based on this, a deep stacked autoencoder-based customer behavior classification model is presented in this research work along with the dragonfly optimization algorithm. The network parameters of the deep-stacked autoencoder are optimized using the dragonfly optimization algorithm to attain enhanced classification accuracy. Benchmark customer behavior dataset is used for experimentation and analyzed the performance in terms of recall, precision, f1-score, and accuracy. The proposed optimized deep learning model attains better performance compared to deep learning approaches like Long-Short Term Memory (LSTM), Convolutional Neural network (CNN), and Autoencoder models. |
| Druh dokumentu: | Article |
| Jazyk: | English |
| ISSN: | 2788-7669 2789-1801 |
| DOI: | 10.53759/7669/jmc202505173 |
| Přístupové číslo: | edsair.doi...........bb2ddfce70a79fecc2169ba65129528e |
| Databáze: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | Customer Relationship Management (CRM) plays a major role in analyzing customer behavior and the opinions of an organization or enterprise. Data mining methods are widely uutilize to analyze customer data to increase business and revenue. Data mining refers to the extraction of essential and useful information from customer feedback and activities on websites through mining technologies. However, extracting essential information from customer behavior is quite challenging as it requires a detailed analysis of customer desires, requirements, buying patterns, etc., all the information in the e-commerce market is essential for an enterprise as it will support knowing the customer behavior. Deep learning algorithms based on customer behavior classification models are evolved in recent times. However, the performance can be improved if the network parameters are optimized through optimization algorithms. Based on this, a deep stacked autoencoder-based customer behavior classification model is presented in this research work along with the dragonfly optimization algorithm. The network parameters of the deep-stacked autoencoder are optimized using the dragonfly optimization algorithm to attain enhanced classification accuracy. Benchmark customer behavior dataset is used for experimentation and analyzed the performance in terms of recall, precision, f1-score, and accuracy. The proposed optimized deep learning model attains better performance compared to deep learning approaches like Long-Short Term Memory (LSTM), Convolutional Neural network (CNN), and Autoencoder models. |
|---|---|
| ISSN: | 27887669 27891801 |
| DOI: | 10.53759/7669/jmc202505173 |