What is a feature in QuantiX? A feature is anything that provides information about candles based on the candle itself and previous candles, such as Open, High, Low, Close, Volume, and indicators. Targets are not features because they use data from subsequent candles. What are the types of features in QuantiX? There are two types of features: numeric, which have numerical values like RSI and MACD, and categorical, which have labels like 'rsi_strategy' and 'k_d_cross'. What is a target in QuantiX? A target is a function that uses future candle data to return labels or numeric values that help quantify future asset performance. Targets cannot be used directly in trading, but can be used to train machine learning models or perform statistical analysis. What are the types of targets in QuantiX? There are two types of targets: numeric targets that assign numerical values to candles and categorical targets that assign labels like Buy, Sell, and No-Action to candles. What is a signal in QuantiX? A signal is a set of conditions combined with logical operands used to trigger buy or sell orders in backtests and bots. Signals are created by combining labels from features using AND and OR operands.

What is Feature Analysis in Quantix? Feature analysis allows you to investigate numeric and categorical features with numeric and categorical targets. How do I access Feature Analysis? Feature analysis can be found under Data Analysis --> Feature Analysis. What types of reports are generated for numeric targets in Feature Analysis? For numeric targets, when one numeric feature is included, the report is a histogram. When two numeric features are included, the report is a grid showing mean target values. When one categorical feature is used, the report is a bar chart showing the mean value of the target for each label. When two categorical features are used, the report is a grid showing mean values for all combinations of labels. When one numeric and one categorical feature are used, the X axis shows the numeric feature, and the Y axis shows the labels of the categorical feature. What types of reports are generated for categorical targets in Feature Analysis? When one numeric feature is used with a categorical target, the report shows the precision of different ranges of the feature for the target's labels. When two numeric features are used, graphs are created showing the precisions of classes for ranges of each feature. When one categorical feature is used, two graphs are reported showing the precision and number of occurrences for classes, and a bar chart of precisions. When multiple categorical features are used the output is an interactive chart with different combinations of features. When one numeric and one categorical feature are used, the output is the same as two numeric features, with the X axis showing the numeric feature and the Y axis showing labels of the categorical feature.

What is the purpose of the Machine Learning section in Quantix? The Machine Learning section provides tools and utilities to use machine learning for strategy development. How do I create a new machine learning model? You can create a new model by navigating to Data Analysis --> ML --> New Model. What are the key steps in training a machine learning model? The key steps are: defining the Train Data, setting Model Parameters, and configuring the Feature Processor. What kind of data is needed to train a machine learning model? To train a model, you need a market data containing numeric features as inputs, and a target that the model will learn to predict. What is 'Model Memory' in the Feature Processor? Memory defines the number of recent candles used as input for the model, with the default being one. Increasing memory can provide more information but may not always improve results. What is the purpose of the Signal Generator in the Machine Learning section? The Signal Generator is used to create buy, sell, and no-action signals from a trained machine learning model. How do I assess the performance of a trained machine learning model? You can evaluate model performance through the Report section, which provides metrics like AUC scores and visualizations of probability distributions.