commit 0c73063754dc61c5da6978a6b78bb8dfc027ddc1 Author: 202410715227 SAVAREL PRAMUDYA HANDOKO <202410715227@mhs.ubharajaya.ac.id> Date: Sat Nov 7 14:49:05 2026 +0700 Upload files to "/" diff --git a/SIPADI_Sistem_Pakar_Tanaman_Padi_Google_Colab.ipynb b/SIPADI_Sistem_Pakar_Tanaman_Padi_Google_Colab.ipynb new file mode 100644 index 0000000..a520784 --- /dev/null +++ b/SIPADI_Sistem_Pakar_Tanaman_Padi_Google_Colab.ipynb @@ -0,0 +1,441 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "93da403c", + "metadata": {}, + "source": [ + "# SIPADI – Sistem Pakar Diagnosis Hama dan Penyakit Tanaman Padi\n", + "\n", + "**Metode:** Rule-Based, Forward Chaining, dan Certainty Factor (CF) \n", + "**Platform:** Google Colab / Python\n", + "\n", + "Notebook ini merupakan prototipe sistem pakar berdasarkan rancangan proposal SIPADI. Pengguna memilih gejala yang terlihat pada tanaman padi, kemudian sistem menelusuri aturan IF–THEN dan menampilkan kandidat diagnosis beserta nilai keyakinannya.\n", + "\n", + "> **Catatan penting:** Daftar gejala, aturan, dan bobot CF di bawah merupakan data awal/ilustratif dari proposal, bukan hasil validasi pakar. Hasil hanya untuk bantuan identifikasi awal, bukan pengganti pemeriksaan lapangan atau konsultasi penyuluh pertanian.\n" + ] + }, + { + "cell_type": "markdown", + "id": "7c656b15", + "metadata": {}, + "source": [ + "## 1. Import library\n", + "\n", + "Library yang digunakan adalah `pandas` untuk menampilkan hasil dan `ipywidgets` untuk membuat formulir konsultasi interaktif di Google Colab.\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "01450574", + "metadata": {}, + "outputs": [], + "source": [ + "import pandas as pd\n", + "from IPython.display import display, clear_output\n", + "import ipywidgets as widgets\n", + "\n", + "print(\"SIPADI siap dijalankan.\")\n" + ] + }, + { + "cell_type": "markdown", + "id": "494e2144", + "metadata": {}, + "source": [ + "## 2. Basis pengetahuan: daftar gejala\n", + "\n", + "Kode gejala mengikuti rancangan proposal SIPADI. Gejala tambahan untuk walang sangit ditanyakan secara terpisah karena proposal mensyaratkan adanya serangga pada bulir.\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "80ffb6be", + "metadata": {}, + "outputs": [], + "source": [ + "GEJALA = {\n", + " \"G01\": \"Daun/tanaman menguning dan pertumbuhan terhambat\",\n", + " \"G02\": \"Petakan tanaman mengering seperti terbakar\",\n", + " \"G03\": \"Ditemukan wereng kecil berwarna cokelat di pangkal batang\",\n", + " \"G04\": \"Anakan atau batang mati dan mudah dicabut\",\n", + " \"G05\": \"Terdapat lubang gerekan pada batang atau malai putih/hampa\",\n", + " \"G06\": \"Terdapat kelompok telur atau larva pada tanaman\",\n", + " \"G07\": \"Bulir berubah warna/cokelat kehitaman dan hampa\",\n", + " \"G08\": \"Bercak berbentuk belah ketupat dengan pusat keabu-abuan pada daun\",\n", + " \"G09\": \"Bercak memanjang hijau keabu-abuan hingga putih dengan tepi tidak beraturan\",\n", + " \"G10\": \"Daun menguning-oranye, tanaman kerdil, anakan berkurang\",\n", + " \"G11\": \"Bercak daun bulat/oval berwarna cokelat dengan pusat lebih pucat\",\n", + " \"G12\": \"Gejala menyebar pada kondisi lembap/hujan\",\n", + " \"G13\": \"Daun menggulung/terlipat memanjang dan tampak bekas gerekan\",\n", + " \"G14\": \"Terdapat bekas gigitan atau tanaman rebah/terpotong\"\n", + "}\n", + "\n", + "df_gejala = pd.DataFrame(\n", + " [{\"Kode\": kode, \"Gejala\": nama} for kode, nama in GEJALA.items()]\n", + ")\n", + "display(df_gejala)\n" + ] + }, + { + "cell_type": "markdown", + "id": "7d68acde", + "metadata": {}, + "source": [ + "## 3. Basis pengetahuan: diagnosis, aturan, dan rekomendasi\n", + "\n", + "Setiap aturan memiliki premis gejala, kode diagnosis, bobot CF pakar, dan rekomendasi awal. Bobot CF di sini hanya contoh untuk demonstrasi program dan harus divalidasi oleh pakar pertanian sebelum penggunaan nyata.\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "cd211936", + "metadata": {}, + "outputs": [], + "source": [ + "DIAGNOSIS = {\n", + " \"P01\": {\n", + " \"nama\": \"Wereng batang cokelat\",\n", + " \"jenis\": \"Hama\",\n", + " \"saran\": \"Lakukan pengamatan rutin pada pangkal rumpun, jaga keseimbangan ekosistem sawah, dan konsultasikan tindakan PHT dengan penyuluh setempat.\"\n", + " },\n", + " \"P02\": {\n", + " \"nama\": \"Penggerek batang\",\n", + " \"jenis\": \"Hama\",\n", + " \"saran\": \"Amati gejala sundep/beluk dan keberadaan kelompok telur. Lakukan pemantauan serta pengendalian sesuai prinsip PHT.\"\n", + " },\n", + " \"P03\": {\n", + " \"nama\": \"Walang sangit\",\n", + " \"jenis\": \"Hama\",\n", + " \"saran\": \"Pantau keberadaan walang sangit pada fase berbunga hingga pengisian bulir. Lakukan pengendalian terpadu sesuai kondisi lapangan.\"\n", + " },\n", + " \"P04\": {\n", + " \"nama\": \"Penyakit blas\",\n", + " \"jenis\": \"Penyakit\",\n", + " \"saran\": \"Gunakan benih sehat, lakukan pemupukan berimbang, dan konsultasikan pengelolaan penyakit dengan penyuluh.\"\n", + " },\n", + " \"P05\": {\n", + " \"nama\": \"Hawar daun bakteri\",\n", + " \"jenis\": \"Penyakit\",\n", + " \"saran\": \"Gunakan varietas tahan bila tersedia, jaga sanitasi lahan, dan hindari pemupukan nitrogen berlebihan.\"\n", + " },\n", + " \"P06\": {\n", + " \"nama\": \"Tungro\",\n", + " \"jenis\": \"Penyakit\",\n", + " \"saran\": \"Cabut dan musnahkan tanaman terinfeksi sesuai arahan petugas, pantau wereng hijau sebagai vektor, dan gunakan varietas tahan.\"\n", + " },\n", + " \"P07\": {\n", + " \"nama\": \"Bercak cokelat\",\n", + " \"jenis\": \"Penyakit\",\n", + " \"saran\": \"Perhatikan kesuburan tanah, gunakan benih sehat, dan lakukan pengelolaan air serta pemupukan yang baik.\"\n", + " },\n", + " \"P08\": {\n", + " \"nama\": \"Tikus sawah\",\n", + " \"jenis\": \"Hama\",\n", + " \"saran\": \"Lakukan sanitasi habitat, pengemposan atau cara pengendalian yang direkomendasikan petugas, serta pengendalian serempak berbasis kelompok.\"\n", + " }\n", + "}\n", + "\n", + "# Format aturan:\n", + "# premis = seluruh gejala yang harus dipilih (AND)\n", + "# cf_pakar = bobot keyakinan pakar ilustratif (0 sampai 1)\n", + "ATURAN = [\n", + " {\"kode\": \"R01\", \"diagnosis\": \"P01\", \"premis\": [\"G02\", \"G03\"], \"cf_pakar\": 0.85},\n", + " {\"kode\": \"R02\", \"diagnosis\": \"P02\", \"premis\": [\"G04\", \"G05\"], \"cf_pakar\": 0.85},\n", + " {\"kode\": \"R03\", \"diagnosis\": \"P03\", \"premis\": [\"G07\"], \"cf_pakar\": 0.70, \"syarat_walang\": True},\n", + " {\"kode\": \"R04\", \"diagnosis\": \"P04\", \"premis\": [\"G08\"], \"cf_pakar\": 0.80},\n", + " {\"kode\": \"R05\", \"diagnosis\": \"P05\", \"premis\": [\"G09\"], \"cf_pakar\": 0.80},\n", + " {\"kode\": \"R06\", \"diagnosis\": \"P06\", \"premis\": [\"G01\", \"G10\"], \"cf_pakar\": 0.85},\n", + " {\"kode\": \"R07\", \"diagnosis\": \"P07\", \"premis\": [\"G11\"], \"cf_pakar\": 0.75},\n", + " {\"kode\": \"R08\", \"diagnosis\": \"P08\", \"premis\": [\"G14\"], \"cf_pakar\": 0.80}\n", + "]\n", + "\n", + "print(\"Jumlah gejala:\", len(GEJALA))\n", + "print(\"Jumlah diagnosis:\", len(DIAGNOSIS))\n", + "print(\"Jumlah aturan:\", len(ATURAN))\n" + ] + }, + { + "cell_type": "markdown", + "id": "21550358", + "metadata": {}, + "source": [ + "## 4. Mesin inferensi Forward Chaining dan Certainty Factor\n", + "\n", + "Forward chaining memeriksa apakah seluruh premis suatu aturan terdapat pada fakta yang dipilih pengguna. Jika terpenuhi, aturan menghasilkan kandidat diagnosis.\n", + "\n", + "Untuk setiap aturan:\n", + "- `CF evidence = CF pakar × CF pengguna`\n", + "- Jika beberapa aturan mendukung diagnosis yang sama, nilai CF positif digabungkan dengan rumus `CFgabung = CF1 + CF2 × (1 − CF1)`.\n", + "\n", + "Pada formulir demo, tingkat keyakinan pengguna dipilih untuk semua gejala sekaligus. Ini merupakan penyederhanaan antarmuka; sistem produksi sebaiknya meminta tingkat keyakinan per gejala.\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "4fdfc46d", + "metadata": {}, + "outputs": [], + "source": [ + "def gabungkan_cf(cf1, cf2):\n", + " \"\"\"Menggabungkan dua nilai CF positif.\"\"\"\n", + " return cf1 + cf2 * (1 - cf1)\n", + "\n", + "\n", + "def forward_chaining_cf(gejala_dipilih, cf_pengguna=1.0, ada_walang=False):\n", + " \"\"\"Menjalankan forward chaining dan menghitung CF setiap kandidat.\"\"\"\n", + " fakta = set(gejala_dipilih)\n", + " hasil = {}\n", + " aturan_aktif = []\n", + "\n", + " for aturan in ATURAN:\n", + " # Semua premis aturan harus menjadi fakta\n", + " if not set(aturan[\"premis\"]).issubset(fakta):\n", + " continue\n", + "\n", + " # Aturan walang sangit membutuhkan konfirmasi serangga pada bulir\n", + " if aturan.get(\"syarat_walang\", False) and not ada_walang:\n", + " continue\n", + "\n", + " kode_diag = aturan[\"diagnosis\"]\n", + " cf_evidence = aturan[\"cf_pakar\"] * cf_pengguna\n", + "\n", + " if kode_diag not in hasil:\n", + " hasil[kode_diag] = {\n", + " \"cf\": cf_evidence,\n", + " \"aturan\": [aturan[\"kode\"]],\n", + " \"gejala_pendukung\": list(aturan[\"premis\"])\n", + " }\n", + " else:\n", + " hasil[kode_diag][\"cf\"] = gabungkan_cf(\n", + " hasil[kode_diag][\"cf\"], cf_evidence\n", + " )\n", + " hasil[kode_diag][\"aturan\"].append(aturan[\"kode\"])\n", + " hasil[kode_diag][\"gejala_pendukung\"].extend(aturan[\"premis\"])\n", + "\n", + " aturan_aktif.append(aturan[\"kode\"])\n", + "\n", + " # Urutkan kandidat dari CF tertinggi\n", + " hasil = dict(sorted(hasil.items(), key=lambda item: item[1][\"cf\"], reverse=True))\n", + " return hasil, aturan_aktif\n", + "\n", + "\n", + "def kategori_keyakinan(cf):\n", + " persen = cf * 100\n", + " if persen >= 80:\n", + " return \"Tinggi\"\n", + " elif persen >= 50:\n", + " return \"Sedang\"\n", + " return \"Rendah\"\n" + ] + }, + { + "cell_type": "markdown", + "id": "6c9f8fd2", + "metadata": {}, + "source": [ + "## 5. Formulir konsultasi interaktif\n", + "\n", + "Centang gejala yang terlihat pada tanaman padi. Pilih juga tingkat keyakinan terhadap gejala yang dipilih, kemudian tekan tombol **Proses Diagnosis**.\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "9ac2bc08", + "metadata": {}, + "outputs": [], + "source": [ + "pilihan_gejala = {\n", + " kode: widgets.Checkbox(value=False, description=f\"{kode} — {nama}\", indent=False,\n", + " layout=widgets.Layout(width=\"100%\"))\n", + " for kode, nama in GEJALA.items()\n", + "}\n", + "\n", + "checkboxes = widgets.VBox(list(pilihan_gejala.values()),\n", + " layout=widgets.Layout(width=\"100%\"))\n", + "\n", + "keyakinan = widgets.Dropdown(\n", + " options=[\n", + " (\"Yakin (1.00)\", 1.00),\n", + " (\"Cukup yakin (0.80)\", 0.80),\n", + " (\"Kurang yakin (0.60)\", 0.60),\n", + " (\"Tidak yakin (0.40)\", 0.40)\n", + " ],\n", + " value=1.00,\n", + " description=\"Keyakinan:\",\n", + " style={\"description_width\": \"initial\"}\n", + ")\n", + "\n", + "konfirmasi_walang = widgets.Checkbox(\n", + " value=False,\n", + " description=\"Saya juga menemukan serangga walang sangit pada bulir padi\",\n", + " indent=False,\n", + " layout=widgets.Layout(width=\"100%\")\n", + ")\n", + "\n", + "tombol = widgets.Button(\n", + " description=\"Proses Diagnosis\",\n", + " button_style=\"success\",\n", + " icon=\"search\"\n", + ")\n", + "\n", + "keluaran = widgets.Output()\n", + "\n", + "def jalankan_konsultasi(_):\n", + " with keluaran:\n", + " clear_output()\n", + " terpilih = [kode for kode, cb in pilihan_gejala.items() if cb.value]\n", + "\n", + " if not terpilih:\n", + " print(\"Silakan pilih minimal satu gejala terlebih dahulu.\")\n", + " return\n", + "\n", + " hasil, aturan_aktif = forward_chaining_cf(\n", + " terpilih,\n", + " cf_pengguna=keyakinan.value,\n", + " ada_walang=konfirmasi_walang.value\n", + " )\n", + "\n", + " print(\"HASIL KONSULTASI SIPADI\")\n", + " print(\"=\" * 55)\n", + " print(\"Gejala yang dipilih:\")\n", + " for kode in terpilih:\n", + " print(f\"- {kode}: {GEJALA[kode]}\")\n", + "\n", + " if not hasil:\n", + " print(\"\\nBelum ditemukan aturan yang cocok dengan kombinasi gejala.\")\n", + " print(\"Tambahkan informasi gejala yang relevan atau konsultasikan langsung\")\n", + " print(\"dengan penyuluh/pakar pertanian.\")\n", + " return\n", + "\n", + " print(\"\\nAturan aktif:\", \", \".join(aturan_aktif))\n", + " print(\"\\nKANDIDAT DIAGNOSIS\")\n", + " baris = []\n", + " for peringkat, (kode, data) in enumerate(hasil.items(), start=1):\n", + " info = DIAGNOSIS[kode]\n", + " baris.append({\n", + " \"Peringkat\": peringkat,\n", + " \"Kode\": kode,\n", + " \"Diagnosis\": info[\"nama\"],\n", + " \"Jenis\": info[\"jenis\"],\n", + " \"CF (%)\": round(data[\"cf\"] * 100, 2),\n", + " \"Kategori\": kategori_keyakinan(data[\"cf\"]),\n", + " \"Aturan aktif\": \", \".join(data[\"aturan\"])\n", + " })\n", + "\n", + " display(pd.DataFrame(baris))\n", + "\n", + " kode_teratas = next(iter(hasil))\n", + " info = DIAGNOSIS[kode_teratas]\n", + " data = hasil[kode_teratas]\n", + "\n", + " print(\"\\nPENJELASAN KANDIDAT TERATAS\")\n", + " print(\"Diagnosis:\", info[\"nama\"])\n", + " print(\"Gejala pendukung:\")\n", + " for kode in sorted(set(data[\"gejala_pendukung\"])):\n", + " print(f\"- {kode}: {GEJALA[kode]}\")\n", + " print(\"\\nSaran penanganan awal:\")\n", + " print(info[\"saran\"])\n", + "\n", + " print(\"\\nPERINGATAN:\")\n", + " print(\"Hasil ini adalah indikasi awal berdasarkan aturan dan bobot ilustratif.\")\n", + " print(\"Nilai CF bukan probabilitas statistik dan belum divalidasi oleh pakar.\")\n", + " print(\"Jangan menggunakan pestisida tanpa mengikuti label dan arahan petugas.\")\n", + "\n", + "tombol.on_click(jalankan_konsultasi)\n", + "\n", + "display(widgets.HTML(\"