Create Stock-Price-Prediction.jl

Stock-Price-Prediction.jl

@@ -0,0 +1,77 @@
+using Flux
+using CSV
+using DataFrames
+using Plots
+using Statistics
+
+# Load and preprocess the dataset
+data_file = "stock_prices.csv"  # Replace with your CSV file containing stock prices
+data = CSV.read(data_file, DataFrame)
+
+# Assuming the CSV has a column "Close" for closing prices
+prices = data.Close
+
+# Normalize the data
+function normalize(data)
+    return (data .- mean(data)) ./ std(data)
+end
+
+normalized_prices = normalize(prices)
+
+# Prepare the dataset for training
+function prepare_data(data, seq_len)
+    X, Y = [], []
+    for i in 1:(length(data) - seq_len)
+        push!(X, data[i:i+seq_len-1])
+        push!(Y, data[i+seq_len])
+    end
+    return hcat(X...)', hcat(Y...)
+end
+
+seq_len = 50  # Sequence length for LSTM
+X, Y = prepare_data(normalized_prices, seq_len)
+
+# Split into training and test sets
+train_ratio = 0.8
+train_size = Int(floor(train_ratio * size(X, 1)))
+
+X_train, X_test = X[1:train_size, :], X[train_size+1:end, :]
+Y_train, Y_test = Y[1:train_size], Y[train_size+1:end]
+
+# Convert to Flux tensors
+X_train = Flux.unsqueeze(Float32.(X_train), dims=3)
+X_test = Flux.unsqueeze(Float32.(X_test), dims=3)
+Y_train = Float32.(Y_train)
+Y_test = Float32.(Y_test)
+
+# Define the LSTM model
+model = Chain(
+    LSTM(seq_len, 50),
+    Dense(50, 1)
+)
+
+# Define loss function and optimizer
+loss(x, y) = Flux.mse(model(x), y)
+opt = ADAM(0.001)
+
+# Training the model
+function train!(epochs)
+    for epoch in 1:epochs
+        for i in 1:train_size
+            Flux.train!(loss, params(model), [(X_train[:, :, i], Y_train[i])], opt)
+        end
+        println("Epoch $epoch: Training Loss = $(loss(X_train, Y_train))")
+    end
+end
+
+train!(50)  # Train for 50 epochs
+
+# Test the model
+predictions = model(X_test)
+
+# Plot the results
+plot(1:length(Y_test), Y_test, label="Actual")
+plot!(1:length(predictions), predictions, label="Predicted", linestyle=:dash)
+
+# Save the model
+Flux.save("stock_price_predictor.bson", model)