IJEAI 

Volume 5 Issue 1


1: Digital Marketing Management and Business Strategy

ABSTRACT: 



The internet brought about disruptive change in the business landscape by spawning a slew of digital marketing tactics. However, with these new tactical options, marketing managers must decide which digital marketing tactics to invest in and prioritize what they want to achieve. Consider these issues through the eyes of four different business strategies:

low-cost defenders, prospectors, differentiated defenders, and analyzers. This article gives marketing managers insight into how businesses pursuing different strategies approach these digital marketing issues, with the ultimate goal of assisting managers in the efficient and effective implementation of their firm's adopted strategy.


Keywords: SEO, Digital marketing strategy, Content Marketing, Marketing campaign.



2: Pay Attention to Correct Region for Semantically Grounded Image Captioning


ABSTRACT: 


Many recent researches have focused on using the attention mechanism and transformer in image captioning tasks to allow the captioning model to align image regions to caption words dynamically, which improved the accuracy of the results. However, most image captioning models have not optimally utilized the attention mechanism, where they relied on linking each generated word to specific regions of an image. This doesn’t always ensure that the generated words are associated with the correct regions because some words are unrelated to the image. Therefore, we proposed an approach to improve the attention mechanism and make it more semantic and meaningful. In our proposed approach, which we named it Pay Attention to Correct Region (PACR), we divided the captioning process into two stages: The first stage is to generate texts that are related to the scene in the image, and the second stage is to generate texts related to language and writing rules, resulting in a final caption that is correct in term of language and meaningful in visual grounding. Our proposed model significantly improved accuracy when generating image

descriptions and was more adaptable to other vision-language tasks. We evaluated our results using Bleu-4, CIDEr, and METEOR scales and obtained 40.3, 133.7, and 30.6 results, respectively.


Keywords: Image Captioning, Transformer, Semantic Attention, Visual Grounding, Vision-Language



3: POSITION FALSIFICATION DETECTION IN VANETS BYN USING MACHINE LEARNING


ABSTRACT: 



Vehicular ad hoc networks (VANETs) have faced privacy and security challenges since their inception starting with communication, routing and various forms of security attacks, which end with the privacy of vehicle information and may threaten driver safety. One of the major issues in VANET security is position falsification misbehavior, which malicious actors could exploit to cause an accident in the path of moving vehicles that would appear to be caused by software or human error without the driver's knowledge and direct traffic in areas where the attackers wish to do so.

These challenges require an IDS system that can accurately predict whether the given information is true or false depend on machine learning algorithms using some techniques to enhance the prediction by adding or removing attributes in order to create a more accurate system. This paper presents novel approaches for assessing and optimizing intrusion distributed and centralized detection systems (IDS) by applying Gradient Boosting algorithms and adding or eliminating features from the dataset using 2BSM public dataset to make the models more applicable of detection this type of security attacks.


Keywords: Ad hoc Networks, Machine Learning, Detection Systems, Certificate, Blacklist, Euclidean Distance.



4: Hiding Data Using Combined Least and Most Significant Bits



ABSTRACT: 


Steganography involves hiding data within an image using a secret key. This study introduces two techniques that utilize both the most significant bit (MSB) and the least significant bit (LSB) in a color image (24-bit RGB). The novel approach combines LSB and MSB bits by checking MSB values and replacing LSB bits with a secret message. The method maintains stego-image quality by ensuring a histogram match between the cover and stego images, enhancing security by not hiding information across the entire image. Evaluation metrics include Mean Square Error (MSE), Payload Computation, and Peak Signal-to-Noise Ratio (PSNR), with high PSNR and low MSE values. Results show a PSNR of 87.141 and MSE of 0.00012 for 80-bit messages, and a PSNR of 72.023 and MSE of 0.0040 for 1200-bit messages. Entropy remains consistent between cover and stego images, indicating enhanced security against attacks.

Keyword: Steganography, LSB. MSB, hiding, PSNR, MSE, image, stegoimage, coverimage, RGB, extracting.

5: A FUZZY BASED POWER GENERATOR DIAGNOSTIC SYSTEM 


ABSTRACT: 


This paper aims to eliminate the difficulties encountered in the diagnosis of faults in power generators, which have led to the adoption of various contemporary approaches to replace older and sometimes trial and error approaches adopted by some generator technicians. Most existing methods do not accurately diagnose these faults and these have led to waste of time and other resources due to faulty or delayed diagnoses hence the need for a more precise approach.

In this research, an approach that uses fuzzy logic based diagnostic system for power generators was adopted. The necessary linguistic variables and the membership functions for the variables used in converting them from crisp data into fuzzy data sets were defined. Finally, an inference engine was created and conversion of output data into non fuzzy values (defuzzification) was done. Simulation and Implementation of the generator diagnostic system was done with MATLAB and Python Programming language from the created predictors using linguistic variables. The result shows that this approach makes it easier to diagnose fault in power generators and with more precision than existing

ones.


Keywords: Generators, Fault Diagnoses, Fuzzy Logic



6: Study the effect of nanoparticles in aluminum alloys processed on mechanical properties


ABSTRACT: 


The research will focus on optimizing and characterizing the Al6061 nanocomposite made using the stir casting technique. The Al6061 matrix was used to distribute and refine the alumina nanoparticles. The different concentrations of the Al2O3 particles were studied. Scanning electron microscopy was used to examine the microstructure of Al6061.

The three main properties of Al6061: hardness, tensile, and ductility were analyzed. Tribological properties of the Al6061 were also studied. The hardness test results revealed that the Al6061 nanocomposite exhibited a significant increase in hardness from 22 to 36BHN at 10-wt% alumina. The improvement ratios were 35.6%, 27.7%, and 34%. It was also observed that the use of Nano reinforcement could improve the wear rate. The study revealed that the Al2O3 nanoparticles with a 0.65 10-3 (mm3/m) load exhibited a better wear rate than those with a 1.22 10-3 (mm3/m) load.


Keywords: nanoparticles, aluminum alloys, mechanical, properties