| Sumario: | In the present world, women harassment and its associated threatening are increasing day-by-day on several places. Specifically, the social mediums are the major reason of such things happening in global manner. The person sitting in front of system can do anything without mentioning any identities, so that it is easy for threatening and do unwanted illegal activities accordingly. The basic motto of this paper is to design a new methodology to provide a safety mechanism to women to prevent themselves from harmful persons. In this paper a new methodology is introduced, which is called as Machine Learning based Social Threatening Filter (MLSTF). This proposed approach identifies the threats raised against women in Social Medias such as FaceBook, Twitter and so on. The logic of machine learning is associated with the proposed approach to identify the tweets raised in social media sites, manage that as a dataset for identifying the harmful contents and find out the word intensities to recognize the contents presented in it. This paper also provides a way to collect the real-time data from the women by means of placing the Smart Wearable Device (SWD) on respective person end to monitor the details instantly. This device contains several intelligent sensor units such as Body Position Identifier, Location, Accident Detection Sensor, Small Mic for recognize the emergency voice to act like a black-box and the pen hole camera for capturing the respective place picture. These all sensors are integrated together into the SWD to monitor the real-time women data and pass such data to the server end for processing by using Internet of Things (IoT). The logic of IoT is used to establish the connection between client end SWD and the remote server end by means of associated internet services. By using this facility the locally accumulated data will be passed to the remote server end for processing, in which the machine learning based training principles verifies the data and find out the associated threats to take the necessary actions accordingly. The process of machine learning accumulates the real-time data from social media Kaggle dataset and train the machine accordingly based on the metrics given in the dataset. The realtime data accumulated from the women end SWD is the testing input to verify the threat content, if so the alert will immediately be raised to the respective individuals for taking appropriate actions. The sentiment analysis process also taken care with the proposed approach of data filtering, in which it analyze the sentiment details based on the activities of the women as well as the tweets raised in social media. In which it can also be analyzed with the help of proposed approach. The proposed approach assures the resulting accuracy level by means of performance level improvements such as data collection accuracy, reduced error rate during data processing, reduced time complexity levels as well as request and response counts. For all the proposed approach of MLSTF is the suitable mechanism to act against women harassments and provides a proper safety metrics to them in an intelligent way as well as the resulting section assures the accuracy levels in good manner.
|
|---|