ShieldSquare named a Leader in The Forrester New Wave™: Bot Management Report, Q3 2018. Click to know more.

ShieldSquare named a Leader in The Forrester New Wave™: Bot Management Report, Q3 2018. Click to know more.


Intent Behind Attacks

Intent-based Deep Behavior Analysis (IDBA)

Intent-based Deep Behavior Analysis (IDBA) performs behavioral analysis at a higher level of abstraction of ‘intent’ unlike the commonly used shallow ‘interaction’- based behavior analysis. IDBA consists of three stages: intent encoding, intent analysis, and adaptive learning. During adaptive learning, we also apply challenge-response authentication that helps us dynamically improve our machine learning models.

Capturing intent enables IDBA to provide significantly higher levels of accuracy while detecting bots with advanced human-like interaction capabilities. IDBA builds upon ShieldSquare’s research findings in semi-supervised machine learning and leverages the latest developments in deep learning.

Unique device fingerprinting for each device

Device fingerprinting is the measurement of browser/ mobile, software, and connection attributes in order to generate a risk profile of a device in real-time. ShieldSquare collects multiple diverse data inputs from where the page is accessed to compute a unique fingerprint for each device.

Using device fingerprinting, ShieldSquare engine can detect a bot operator’s device even when they change their identities.

unique device finger-print scanning

Bot Identity Dynamic Turning Tests

Dynamic Turing tests to uncover bot identity

When a website or mobile application receives a request to display a page, the embedded API and JS collect multiple parameters (such as browser details) about the accessing entity.

Based on the received data, Turing tests are constructed in real-time to evaluate the accessing entity’s capabilities and behavior to uncover bot identity.

IP tracking tests

ShieldSquare performs network forensics on the received request and identifies if the requests are coming from Tor/ proxy IPs..

Multiple parameters are analyzed including IP address, extracted IP geo-location details, ISP information, IP owner, connection type, etc., to determine if the access is from genuine users or bots.

IP Tracking Test
User Behavior analyzes

User behavior analysis

The behavior of a user on a web page or mobile app will be significantly different from the behavior of an automated bot. Typical users of a website or mobile app have a behavioral characteristic in terms of number of pages visited per session, time spent on each page, frequency of repeat visits, and so on.

A user model is constructed for each individual site based on historical data that can be checked for anomalies and deviation through which bot activity can be unearthed with accuracy.

Collective Bot Intelligence

Collective bot intelligence gathered from all our customers are utilized to identify bots, flag them, and share the intelligence with other websites to ensure that bots and scrapers are identified and necessary actions taken against them.

Data from 3rd party fraud intelligence directories and services are also gathered to keep track of flagged IPs and devices to remediate the impact of malicious bot networks.

Collective Bot Intelligence Report

Machine Learning Bot Detection

Machine learning for efficient bot detection

The ShieldSquare platform uses sophisticated machine learning algorithms that analyze user history, their behavior, and meta-data to accurately and proactively detect and prevent new generations of attacks by malicious bots.

ShieldSquare machine learning algorithms get smarter everyday by learning from new data and feedback that are collected from clients, and identifies emerging bot patterns in real time.

Staying one step ahead of the bots

Our technology is constantly refined by our data science team. Ensuring zero complacency towards bots impacting your business, our bot detection algorithm is tailored by our experts to suit your business and industry specific needs.

Bots are evolving to be more sophisticated and are developing evasion techniques to bypass inspection, even by advanced algorithms. Unsupervised and semi-supervised learning techniques are employed to identify emerging anomalous trends. Our threat research team is focused on improving the accuracy of methodologies to ensure that new bots are caught with zero false positives.

One step-ahead of Bots

Protect your website now

Powered by Think201